Floating Sensorised Networked Robots for Water Monitoring

Final Report Summary - HYDRONET (Floating Sensorised Networked Robots for Water Monitoring)

Executive Summary:
The HydroNet Project is a STREP presented within the ENVIRONMENT thematic area of EC FP7. The main purpose of this document is to present HydroNet context and objectives, to describe its main Scientific and Technological (S&T) results/foregrounds, to overview overcome obstacles and prototypes weakness, to discuss its potential impacts including the socio-economic effects and the wider societal implications, to synthesize the dissemination activities and exploitation of its results.
The HydroNet Project aimed to design, develop and demonstrate a new open hardware (HW) and software (SW) technological platform to improve the monitoring of water bodies (coastal seas, rivers and lakes, possibly with shallow waters) and assess in real time their healthiness by supplying spatial and temporal information. The platform is composed by a network of autonomous, miniaturised, sensorised, radio interconnected, water robots embedded into an Ambient Intelligence (AmI) control suite. The sensorised robot network collects water samples and in-situ analyse several chemical and physical parameters allowing informing in real time, continuously about the well-being status of water bodies.
The platform core is represented by Sensors Devices (biological, optical and electrochemical SDs) hosted inside floating robots and fixed buoys able to communicate with a remote (land) control station. The new miniaturised, original, custom-made SDs detect many pollutants (chromate, cadmium, mercury, dissolved petroleum hydrocarbons and oil slicks) in water sampled up to 50m depth by an electronic syringe sampling subsystem.
The AmI Framework encompasses all the software infrastructure and is responsible of interconnecting the HydroNet robots, human operators and Artificial Intelligence (AI) SW in order to manage and task the fleet, to monitor in real-time all the events occurring at a HydroNet site and to produce valuable information about the observed environment. A radio module allows the data exchange between robots/buoys and the AmI platform. The software on the radio module enables a multihop routing and a SW library provides an easy-to-use send/receive data interface between the computerised network components.
Enhanced mathematical models have been developed to simulate the pollutants biological/chemical processes and transport in rivers, lakes and in coastal waters. New traditional measurements have been performed in the pilot sites (So?a/Isonzo River - Slovenia, Marano Lagoon and Gulf of Trieste - Italy, Livorno coastal area - Italy) to update the scientific environment knowledge and to furnish the needed data to the mathematical dispersion models.
All requirements were identified by the HydroNet Consortium together with Public Institutions and potential end-users of the new platform aiming for a new approach of environmental monitoring with potentially high usability, sustainability and impact. The HydroNet results, lessons learnt, dissemination activities and plans for the HydroNet platform exploitation will be finally illustrated.
Project Context and Objectives:
Water is the most precious resource for the mankind. Water is a limited and vulnerable resource. The use of water affects the quality of this resource itself as well as the nature in a broader sense. Sea, river and lake water is valuable for the environment, irrigation and drinking. It is fundamental to protect and fairly use the water. It is essential to manage the supply and disposal of water wisely to ensure that clean water continues to be available to us and to future generations at an affordable cost. The Directive 2000/60/EC of the European Parliament and of the Council of 23 October 2000, establishing a framework for European Community action in the field of water policy, begins with the statement Water is not a commercial product like any other but, rather, a heritage which must be protected, defended and treated as such. The European Water Framework Directive (WFD) highlighted the need for water quality monitoring in freshwater systems, estuaries and at sea.
Approximately 97% of water in the world is sea water and 3% is fresh water. Of the fresh water only 13% is accessible, which is 0.4% of the total amount of water. Today more than 2 billion people are affected by water shortages in over 40 countries. Two million tons per day of human waste are deposited in water courses. Half the population of the developing world is exposed to polluted sources of water that increase disease incidence. The increase in numbers of people from 6 billion to 9 billion will be the main driver of water resources management for the next 50 years.
All over the world water is in great demand: everybody wants and needs it because it is the basis of life. The World Water Council estimates that domestic water use worldwide will increase by about 40% for the next two decades. Because of the ongoing migration the demand for safe drinking water and proper sanitation is sky rocketing in urban areas where 65% of the world population will live by the year 2035. Some 17% more water is needed to grow food for a growing population. In addition, water demand for industry and energy generation will increase rapidly.
Water is essential for ecosystems and agriculture. Large quantities of water are used for irrigation. Half of the land in the European Union is farmed, mostly by small and medium sized or family run enterprises. Worldwide, agriculture accounts for two thirds of all water used mainly for irrigation.
In Europe, about 30% of the abstracted freshwater is used for agricultural purposes, and up to 75% in Southern Europe. This has a strong impact on nature and water resources. Intensive use of land and water for agriculture significantly contributes to diffuse water pollution through livestock excreta and inappropriate use of fertilisers and pesticides.

Oil hydrocarbon spills can dramatically impact water quality and eco-systems. Despite of the deep environmental impact of dramatic accidents, the spilling due to navigation accidents represents only the 12 % of the total hydrocarbons contaminating the water bodies. The remain is due to less traceable but continuous spills: urban, industrial oil draining and refineries/pipelines losses (37%), tankers routine operations (33%), atmospheric evaporated hydrocarbon fallout (9%), etc.
Summarizing, a large amount of toxic aromatic hydrocarbons is spilled in water bodies every year causing dramatic eco-systems effects. Two visible effects kinds are produced: acute in the short term and chronic in the long period. About the acute effects, the dispersed oil creates a thin film on the water surface avoiding the gaseous exchange and causing conditions of anoxia and clings to the organisms (marine mammals, birds and algae) which compromise their normal life functions. The chronic effects appear later in the living beings when pollutants cause physiological and behavioral alterations. Hydrocarbons in the environment undergo chemo-physical transformations producing CO2 and CO emissions and organic oxygenated compounds. The oil settled on the seafloor is the most harmful remaining unaltered for years and being able to interfere with the life forms. Monitor to rapidly intervene and also to dissuade is the best answer to the oil spills pollution.

New strategies and new radical different approaches are needed to improve the management of water bodies, in terms of:
1. increasing the quality and efficient use of freshwater by developing best practices,
2. reducing the undesirable effects of land use and human activities on water quality,
3. identifying new technologies to assess the chemical and ecological status of water bodies.

Relatively recent advancements in the field of the sensing technologies have brought new trends in the environmental field. The progress in micro-electronics and micro-fabrication technologies has allowed a miniaturization of sensors and devices, opening a series of new exciting possibilities for pollutants monitoring. Moreover, robotics and advanced ICT-based technology (in particular, the extensive use of remote sensing and telemetry) can dramatically improve the detection and prediction of risk/crisis situations related to water pollution, providing new tools for the global management of water resources.

The HydroNet Project, a STREP in the ENVIRONMENT thematic area of EC FP7, aimed at designing, developing and testing an new HW and SW platform composed by a network of autonomous, miniaturised, sensorised, radio interconnected, floating and anchored robots embedded into an Ambient Intelligence (AmI) SW control suite to assess in real-time, in-situ the healthiness of water bodies and to supply information on spatial and temporal water quality.

The core of the HydroNet platform are SDs (bio-, optical- and chemo-SDs) hosted inside the robots able to communicate with the remote (land) AmI control station SW. The movable SDs network is able to analyse several chemical and physical parameters to monitor in real-time, in-situ and continuously the well-being status of water bodies. In real time, the environmental national agencies can know the pollutants in water such as chromate, cadmium, mercury and oil.
Enhanced mathematical models are also developed for simulating the pollutants biological processes and transport in rivers, lakes and in coastal waters preventing wide environmental disaster.

The SDs are embedded into anchored stations (like fixed buoys) and floating robots able to navigate in a network configuration in diverse water scenarios, from coastal sea waters, to rivers (both at the head and mouth of the rivers), to natural and artificial lakes and lagoons. The floating robots navigate rivers, downstream and upstream. The fixed buoys, running as a network node, cooperate with the floating robots to widen the intercommunication and localization among the robots.

The robots are small, lightweight, energy efficient, environmentally compatible (in shape, size, colour, materials and energy supply). The robots communicate with the AmI control station by using a wireless radio connection. The radio communication infrastructure overcomes the required distances without relying on existing devices. The AmI SW provides a friendly user interface to manage the robot network and to present the measured water parameters in real-time. The AmI SW is also connected to an external server for data management and intelligent analysis (databases and data fusion). The AmI SW platform integrates not only sensors and tools for monitoring the environment and robot tasks execution, but also implements communications backhaul systems, databases technologies, knowledge discovery in databases (KDD) processes for extracting and increasing knowledge on water bodies management. Following the computation on stored data, feedback are sent back to human actors (supervisors, decision makers, industrial people, etc.) and/or artificial actuators, in order to perform actions.
A radio module with 2.5W output power on 434MHz with 12V power supply allows the data exchange between robots/buoys and the AmI platform through a RS232 interface. The software on the radio module enables a multihop routing and a SW library provides an easy-to-use send/receive data interface between the computerised network components.

In summary:
- rivers, lakes and coastal waters are vital natural resources, they provide drinking water, crucial habitats for many different types of wildlife and are an important resource for industry and recreation;
- a significant portion of them are environmentally damaged or under threat;
- protecting and improving the environment is an important part of achieving sustainable development and is vital for the long term health, well being and prosperity of everyone.

Traditional environmental monitoring systems are static and their application to dynamic, pervasive monitoring is very expensive. Water cruises with goal of mapping of spatial pollutant distribution are expensive (expenses for employees and exploitation of ships). Application of miniaturised sensorised robots to environmental monitoring should decrease costs of monitoring and give benefits associated with new applications. A network of buoys and floating sensorised mini-robots is autonomous and automatic (without human participation, install&forget) with low power consumption and costs of maintenance. Mini-robots give possibilities to achieve tasks impossible to do with traditional systems (e.g. pollutants sources finder, distributed and pervasive environmental monitoring).

HydroNet platform intends to change drastically the way to monitor the chemical and ecological status of water bodies by introducing the most advanced technological solutions in terms of networks of sensors, robots and remote control that spatially and temporally supply a large variety of useful and novel data on water quality assessment.
In summary, the overall goal of HydroNet is to generate convincing quantitative and qualitative evidence for policy makers and environmental stakeholders demonstrating that the benefits from the HydroNet platform fully justify their implementation.

The main socio-economic objectives of HydroNet are:
1. Determine the basic dimensions of benefits and impact due to the implementation of the HydroNet platform;
2. Develop procedures that can be applied in cost-benefit analysis of environmental monitoring systems with special focus on monitoring heavy metals and oil in the aquatic environment.

Finally, the HydroNet Project aimed at contributing significantly to the success of the new EU Water Framework Directive which is a welcome and radical improvement on earlier, piecemeal EU water legislation. It expands the scope of water protection to all waters and sets out clear objectives that must be achieved by specified dates. The purpose of the Directive is to establish a framework for the protection of inland surface waters (rivers and lakes), transitional waters (estuaries), coastal waters and groundwater. It will ensure all aquatic ecosystems and, with regard to their water needs, terrestrial ecosystems and wetlands meet 'good status' by 2015.
The HydroNet Consortium collaborated with Public Institutions and potential private end-user aiming to propose a new sustainable approach of environmental monitoring with potentially high usability and impact.

The highly interdisciplinary Consortium includes experts in robotics and sensory systems, environmental science scientists (biologists, chemists, modellers, etc.), computer scientists and mathematicians. The project involves ten partners from five European countries: Italy, Norway, Russia, Slovenia, Switzerland and Israel.
Project Results:
The HydroNet Project designed, developed and tested a new platform composed by a network of sensorised, autonomous, cooperating, miniaturised, floating and anchored, robots embedded into an Ambient Intelligence platform to assess in real-time, in-situ the healthiness of water bodies and to supply information on spatial and temporal water quality.

The robots are small, lightweight, energy efficient, environmentally compatible (in shape, size, colour, materials and energy supply). Two floating robot types, optimized for all these different natural scenarios, are available. A multi-hull (catamaran) shape has been chosen for the sea surface robot due to its great capability to perform a safe navigation and to hold a fix position while performing the sampling task. No bulbs or appendixes that may cause problems with sea garbage are present. Low water depths (<1 m), surfacing and emerging obstacles, high maneuverability, long battery endurance, environment and people compliant imposed a different, special boat type for rivers and lagoons: a silent, flat-bottom robot able to cut seaweed. The floating robots are able to navigate in a network configuration in diverse water scenarios, from coastal sea waters, to rivers (both at the head and mouth of the rivers), to natural and artificial lakes and lagoons. The buoys, running as a fixed node of the network, cooperate with the floating robots, to facilitate the communication and localization among the robots and widen the monitorised environmental aquatic area. Special buoys types, atmospheric and electrochemical, also contribute to the water bodies monitoring.

The realized Sensors Devices (SDs) network is the core of the HydroNet Platform and is able to sample and analyse several chemical and physical parameters in water to monitor in real-time, in-situ and continuously the well-being status of water bodies. Chemo- optical- and bio-SDs have been developed and used for the monitoring of physical parameters and pollutants in water such as chromate, cadmium, mercury, oil in and on the water. The SDs are embedded into the fixed stations (like buoys) and the floating robots.

The robots and SDs are part of an Ambient Intelligence (AmI) platform, which integrate not only SDs and tools for monitoring the environment and robot tasks execution, but also communications backhaul systems, databases technologies, knowledge discovery in databases (KDD) processes for extracting and increasing knowledge on water bodies management. Following the computation on stored data, feedbacks are sent back to human actors (supervisors, decision makers, industrial people, etc.) and/or artificial actuators, in order to perform actions. AmI provides a friendly user interface to manage the robot network and to present the measured water parameters in real-time. Moreover, AmI implements an intelligent algorithm called EPSO (Explorative Particle Swarm Optimization) that allows the sensorised robot fleet to identify a pollutant source in a cooperative mode. Finally, enhanced dispersion models simulating pollutants transport and processes in rivers, lakes and in coastal waters have been also integrated into AmI providing tools for simulate the pollutants biological/chemical processes and transport in rivers, lakes and in coastal waters and estimating pollution distribution.

In summary, HydroNet platform allows replacing many actions of the current monitoring procedure by providing immediate results of the status of water body for consultation.
The basic HydroNet platform scenario is the following:
1. a robot starts from the cost (typically from a port) and moves to the sampling point;
2. the robot samples water at -10, -20, -30, -40 and -50 meter, feed the sensor with the samples and analyses the samples;
3. the results of the analysis are sent in real time to the AmI and stored in the database;
4. if requested, the robot moves to the next sampling point and continue at 2;
5. the robot moves back to the port.

By looking to the current campaigns, the robot should perform these operations every three or four months on the average at the same site. Between these samplings at the same location, the robot can be moved to a different area for water monitoring.
To demonstrate the capability of HydroNet to perform typical monitoring tasks, several demonstrations and tests have been performed during the project and even later. A demonstration was held in Livorno on January 28 concurrently with the final project review. In the demonstration the catamaran was equipped with with Cadmium SD, Oil-on-water SD and physical probe,. The flat-boat was equipped with Oil-in-water SD, mercury bioSD and Oil-on-water SD. The Electrochemical Sensor Buoy was equipped with Oil-in-water SD. The demonstration site, shown in Fig. 2, was a part of the Darsena Nuova in the Livorno harbour.

In the Italian final demonstration the following two missions were performed:
First mission:
- AmI commanded Catamaran to move from point A to point B and to sample water at 1 meter depth and distribute water to Cadmium SD;
- Catamaran moved from point A to point B;
- Catamaran sampled water at 1 meter depth and distributed water to Cadmium SD;
- Catamaran analysed the presence of Cadmium in water and sent the result to AmI;
- Ami commanded Catamaran to move from point B to point A;
- Catamaran moved from point B to point A;

Second mission:
- AmI commanded Flat-boat to move from point C to Point D and to sample surface water and distribute water to Oil-in-water SD and Mercury bioSD;
- Flat-boat moved from point C to Point D
- Flat-boat sampled surface water and distributed water to oil-in-water SD and Hg bioSD;
- Flat-Boat analysed presence of oil and mercury in water and sent the results to AmI;
- Ami commanded Flat-Boat to move from point D to point C;
- Flat-boat moved from point D to Point C.

The HUJI Cd(II) measured 4.9 µg/l (low limit 1 µg/l). In the middle of their navigation line, the Lumex CRAB SDs of both the catamaran and the flat-boat measured a value higher than 1200 means which revels the presence of oil on the water. To support this, Lumex dissolved oil and an on ground SD measured 1.15 mg/l showing the presence of hydrocarbons in water. The following table summarizes the measures coming from the physical probe at 0.5m and 1m of depth.

Depth
[m] Temp
[°] Press
[uS/cm] Cond
[uS/cm] Resist
[ohm*cm] TDS
[g/l] Salinity
[ppt] DO
[mg/l] pH ORP
[mV] NO3
[mg/l]
0.5 12.4 61959 45669 22.39 40.27 41.39 12.63 8.08 146.5 0.50
1 12.2 60741 45940 21.77 39.48 40.61 13.06 8.16 157.7 1.04

More tests of the flat-boat have been done on March 13 in front of the SSSA Marine Robotics Laboratory in Livorno. In the tests the flat-boat was equipped with oil-in-water SD (AE-2 Mini SD by Lumex) and oil-on-water SD (CRAB by Lumex) and was commanded with AmI to perform several missions. The robot missions were to perform water sampling and analysis in different points, traveling following simple trajectory or more complex paths. The flat-boat samplings points and trajectories of the missions on March the 13th, 2012 are shown in Fig. 4. The mission results are shortly described in section 4.1.3.1.2 .

A long-duration test has been carried out with the flat-boat on August the 29th, 2012 at Lago Braccini in Pontedera/Italy (see Braccini Lake mission on August 2012 video on the HydroNet WEB site).
The trial goal was to test and validate flat-boat performances during a long-duration and long-range mission. Since lake small dimensions didnt allow to perform a real long rage mission, a shortest path has been continuously carried out more times to simulate a long-range mission.
The single path is shown in Fig. 5 and it is a triangle with vertices P1, P2 and P3: the total triangle length is about 750 meters and the longest side (P3-P1) is 350 meters. The flat-boat was commanded by a remote PC with AmI to move cyclically from P3 to P3 passing from P1 and P2 for 32 times to simulate a total path of about 24 Km. In each waypoint the robot was stopped for about 10 seconds and the commanded speed was 2.2 knots or 1.12 m/s. The robot was equipped with AE-2 Mini and CRAB SDs by Lumex and powered with both batteries and solar panel. Collected measurements and navigation data were stored by the AmI.
Weather conditions were goods: sunny day, temperature between 28 ° and 33° and constant wind from West at 10/15 Kmh. Due to lake small dimensions no waves and strong currents present.
In summary, except for a break of about 30 minutes after cycle #17, the flat-boat travelled continuously from 11:28 in the morning to 17:59 in the afternoon for a total of about 5 hours and 58 minutes. Including three failed cycles (#17, #20 and #28), due to a temporary lack of communication between the ground PC and floating robot, the flat-boat totally travelled almost 22km at the average speed of 1.03 m/s or 2 knots by employing an average of 11 minutes and 12 seconds (including the 10 stop second in each waypoints). By excluding these stop times the average speed was 1.13 m/s or 2.2 knots as programmed.
The Fig. 6 shows all the CRAB sampling points, Table I reports measures and positions of the CRAB SD in round #10.

Sensor name Lat Lon Coeffic.
curvature Surface reflection
LUMSL01 43,659068 10,661774 14 532
LUMSL01 43,659524 10,662491 13 670
LUMSL01 43,660032 10,663178 10 492
LUMSL01 43,660569 10,663774 11 571
LUMSL01 43,661129 10,664172 11 916
LUMSL01 43,661200 10,663945 8 754
LUMSL01 43,660886 10,663115 10 663
LUMOI01 43,660708 10,662231 10 663
LUMSL01 43,660777 10,661752 18 817
LUMSL01 43,660107 10,661532 10 492
LUMSL01 43,659444 10,661427 11 704
LUMSL01 43,658907 10,661353 14 532

Table I - CRAB measures and positions in mission #10.

Official demonstrations showed the catamaran was able to operate in relatively short-term missions. In these demonstration periods, the different subsystems composing the robot were tested and validated. To investigate the robot behavior in a long range cruise (and for a long time period of continuous operations) we carried out a mission along the Livorno coastal area. The robot mission was to perform a sampling in front of Antignano (a village located south of Livorno). This long-range cruise, carried out on 16/3/2012, aimed at different objectives:
? testing the robot system robustness in navigation covering a long distance and implying a long time period of autonomous and continuous operation;
? investigating the energy consumption in a realistic scenario of environmental monitoring.

The mission is showed in Fig. 7. The robot was equipped with one CRAB Lumex SD for oil slick detections, two Lumex Fluorat AE-2 Mini SDs (one fed by one pump for superficial sampling) and one fed by the HydroNet sampling system and with one HUJI Cd SD. The catamaran started the mission in front of the SSSA Marine Robotics Lab and was commanded with several waypoints. The cruise speed was changed to investigate the behavior in all the operative speed range (1.2/1.3/1.4/1.5 m/s).

Depth (m) 0.63 3.59 6.61 9.70 10.68
Temp (C) 13.68 13.63 13.35 13.26 13.27
Cond (mS/cm) 49145.00 49099.00 48768.00 48664.00 48669.00
Tds (g/l) 40.75 40.76 40.77 40.77 40.77
Salinity (ppt) 42.18 42.19 42.18 42.18 42.18
DO (mg/l) 9.20 7.67 7.87 7.83 7.72
pH (units) 8.16 8.17 8.17 8.18 8.18
NO3 (mg/l ) 0.00 0.02 0.01 0.01 0.01
ORP (mV) 217.00 215.50 214.20 213.10 212.10

Depth (m) Measurement value
LUMEX
AE-2 - n°1 0 0.04 mg/l
LUMEX
AE-2 - n°2 11 0.02 mg/l
HUJI Cd 11 0 ppb

The catamaran navigated along the coast with the CRAB measuring every 2 minutes. We report some data on the figure. The measurements presented a low intensity well below the value indicating the presence of oil slicks (1200). In front of Antignano, the catamaran performed a sampling at 11 m. The water was analyzed by one Lumex AE-2 Mini SD and the HUJI Cd SD. Furthermore, a superficial sampling was carried out with the pump and the water was analyzed by the second Lumex AE-2 Mini SD. The previous tables report some data produced by the probe along the water column the results of the Lumex and HUJI SDs.

The robot performed the mission navigating autonomously for 12.562 km. The average surge speed during the mission was about 1.3 m/s. The energy consumption was logged by monitoring the voltage of the battery pack voltage and the supplied current. The overall spent energy was about 820 Wh. Considered the designed battery pack has a total energy capacity of 1800 Wh, this test shows the catamaran can navigate for at least 20 km as required in the design phase.

Lessons learnt and conclusions
The previous described missions show the capability of the HydroNet platform to carry out the intended scenario monitoring. Nevertheless, HydroNet project realized an excellent research prototypal platform, but an industrial prototype is still far away.

By an hard work, the following weaknesses were identified providing new knowledge to our teams:
a. the final floating robots, mainly catamarans, are not robust, reliable and operationally accurate for an extensive use in time and space as required by the state environmental agencies; nowadays, at least 4/5 series each of 10 launches in consecutive days are needed by the environmental agencies,
b. the developed SDs need to be optimized from an electronic/electric point of view; the heavy metal SDs and the bioSD need to be fully validated and certified to be accepted by the environmental agencies for real monitoring campaigns,
c. further SDs, compatible with the characteristics of the platform, should be developed to meet completely the requirements of WFD; floating robots could simply collect water samples and bring them to the ground.
d. measured data should be exported in open, standard formats to be uploaded in world-wide databases made available to the entire scientific community,
e. the user interface should be improved to allow a friendly use by inexpert persons, for example by substituting the current interface navigation oriented with a mission oriented interface,
f. accurate, detailed, zoomable maps available on the professional market should be used to execute samplings very near to sea coastlines or cliffs in a very safe way,
g. the radio communications should be extended to cover an higher zone without the boats use because these latter pose logistic and bureaucratic problems,
h. waters forecast models, timely and spatially very large, should be provided and integrated in the HydroNet platform to allow fruitful comparisons with measured values and effective alerts to be detected by wider HydroNet platform monitoring surprise campaigns,

The HydroNet project developed a new tool for water monitoring operators to perform water sampling and analysis in a more viable and cost effective way and it is intended to partially substitute traditional procedures which currently are performed according to practices described in chapter 4.1.4.1. A significant added value provided by HydroNet is also its computational intelligence in the work-flow of a typical application scenario in order to assist human operators in taking decisions, such as the EPSO algorithm for pollution source finding integrated into AmI.
The floating boats can be used with the maximum designed SDs number and the SW can sustain the maximum platform complexity as shown by the performed demonstrations and long range cruises. Not all SDs were available to perform the HydroNet platform validation, but it has been equipped with the maximum expected components number and it works within limits described below. Since all the interfaces are identical, when they are ready, they can be mounted without altering the platform prototype functionality.

Beside previous technical and methodological HydroNet project outcomes, the Consortium achieved also general socio-economic results, namely:
1. Determine the basic dimensions of benefits and impact due to the implementation of the HydroNet platform,
2. Develop procedures that can be applied in cost-benefit analysis of environmental monitoring systems with special focus on monitoring heavy metals and oil in the aquatic environment.
4.1.3.1. HydroNet Robots
The robots are small, lightweight, floating, energy efficient, environmentally compatible (in shape, size, colour, materials and energy supply) and are able to navigate autonomously and in coordination with each other, in a variety of scenarios, such as coastal waters, artificial and natural lakes, lagoons, rivers. These robots are able to self-orient, to self-organise and to communicate wireless with one another (using state of the art technology in term of wireless communication network) and with an external PC server (e.g. a PC at the control station) for data management and analysis (databases and data fusion), high level network management and high level user interface. The robots host miniaturised, original SDs able to monitor in real-time many different, physical and chemical, environmental parameters, such as chromate, cadmium, mercury, oil, salinity, temperature, chlorophyll, etc.. Particular attention has been devoted to the energy consumption of the robots during the execution of the different tasks.

Two floating robot types, optimized to navigate in different natural water scenarios: coastal sea waters, river mouths, natural and artificial lakes and lagoons are available:
a river robot (flat-boat): small, agile, low payload equipped;
a sea robot (catamaran): small, floating, fully equipped.
The robots main characteristics are:
- minimum range: 15 km for the flat-boat, 20 km for the catamaran;
- 3.5 knots maximum cruise speed;
- 6 hrs continuous operations (minimum);
- sea force 3 (wind speed 7-10 kts, maximum);
- 50 m sampling depth (maximum, for catamarans only);
- reduced dimensions (length < 2 m, < 80 kg) to enable two persons only to manage a robot.
The robots can navigate rivers downstream, stop and keep their position.

HydroNet Robots Main Controller (HRMC), installed on all the robot types, and the original SDs communicate through the CAN protocol with a baudrate of 1Mb and standard CAN messages (11 bits for the identifier). The CAN protocol provides for sending messages in broadcast to all the nodes on the on-board network (see Fig. 8). Each message has an header identifying the message type and the origin/destination device and an 8 bytes payload data.

The communication protocol is divided into two phases: Initialization and Operative.
During the Initialization, each SD is initialized and receives a positional ID (possible pIDs 1--5) corresponding to its position in the floating robot lockers. In this phase, the SD is connected to an external PC only.
Once the Initialization phase is finished, the SD is connected to the robots CAN network and starts working in the Operative phase. During the Operative phase, the HRMC sends commands to any SD by changing the CAN message header. The SD replies by using a header corresponding to its current pID. The whole protocol description and the messages exchanged between the HRMC and SDs are described in the document HydroNet - HRMC-SD Communication Protocol v1.8.pdf (also appended in D5.3 Firmware Data Acquisition and Protocol Data Transfer).

Lessons learnt and conclusions
The floating robots, mainly the two catamarans, strengthen is a general issue concerned with all the mechanical, electrical, electronic and informatics components.

The difficulties arisen from the water, which is the primary element of any HydroNet operative field, should have been more and more carefully considered. It is not enough to design and develop good shapes for an excellent floating robot navigation and good route following algorithms; the water enter everywhere stopping electronics and mechanical components and, at great depths, the pressured water enters in unexpected places. The previous considerations are obvious, but an experienced team should be advised because the experience acquisition has been proved costly and time consuming.

With so many partners cooperating, the interfaces standardisation (electrical, electronic, fluidic, informatics interfaces) among different components should have been undertaken since the project beginning and pursued with determination until the project end. Its surely very hard to execute even
small changes in components when they are ready.
More and more time should have estimated for component integration in laboratory and the test of the complete platform in the field sites (sea, lagoon, river). These field sites require the realization of very solid components and raise a much more difficult work with respect to the calm and clear conditions of laboratory desks. For both reasons the total work time swells.

Another important lesson learnt should be mentioned: the large amount of time needed to launch and put back a floating robot test has not been foreseen. Times to make available the helper raft, to take care of the participants safety, to deploy the robots, to clean and restore everything at the mission completion were initially ignored. This led to unexpected delays in other activities and to a higher, unbudgeted costs. It should not be ignored that the sea/lagoon/rivers conditions often hinder the execution of the field test of this floating robot type and delays of 15 days for a simple component test should be taken in consideration. Because of bad weather conditions, delays easily accumulate.

Procedures and components should be advised to make easier the recovery of the floating robots in malfunction events (for example an automatic back to be control station if the sampling operation or the SDs have failed). Security issues of the floating robots in navigation and in ground movements should be addressed. The obstacle avoidance is a powerful approach which requires, any case, the alarm signal (lights and horn) use in approaching other vessels or persons or grounds.

Improvements are needed for the obstacle detection systems which results weak and unreliable in case of environmental disturbances such as high waves or complex barriers shapes or many closed obstacles. Moreover deep tests should be done on the sonar, its capability and integration into the navigation system. Furthermore, a wind sensor can be useful for improving the navigation system and energy management.

It required huge efforts to transforming high-tech devices (e.g. winch, syringes, fluidic subsystem, ect.) from prototypes working on bench to a reliable monitoring integrated system able to operate in-situ, in the water. This task required deep understanding and management of the physical, dimensional, assembly and power constraints since the design phase.

The different missions carried out in Livorno on March 2012 and fully described in D3.9 showed that the performance navigation, the power consumption, the sampling capabilities of the HydroNet floating boats are coherent with the requirements set at the design phase. It can work for the expected time of an isolate real mission. An one-day monitoring mission similar to that carried out by environmental agencies in terms of covered distance and duration can be executed. The catamarans weakness largely prevent their repetitive, endurance use by state monitoring companies, the flat-boat turned out to be much more robust.

The HydroNet project and the realization of the floating robots allows to SSSA and RT, involved in robotized boats development, to acquire competences and knowledge in marine engineering, almost absent before the project. Main competences gained concern the selection of proper materials for the marine environment, the boat hulls design, the water proof components design also for significant depths and the design and implementation of devices and control strategies for navigation.
4.1.3.1.1. Catamaran robot
The solution adopted for the coastal water scenarios is based on a catamaran.
The HydroNet catamaran, developed by Scuola Superiore SantAnna, is an autonomous surface vehicle (ASV) and consists of two hulls joined by a deck. It is actuated by two independently controlled electric propellers and two rudders, controlled by a single electric motor. The control and navigation system (motor drivers and on-board computers) and the navigation sensors (laser, sonar, altimeter, GPS, digital compass and water speed sensor) allow the robot to move autonomously in its working environment through a series of waypoints set by a ground control station avoiding obstacles. The robot is equipped of up to 4 original SDs (Hg, Cr, Cd, dissolved hydrocarbons), of an outboard SD to detect the water surface oil and of a sampling and fluidic subsystem to collect and distribute water to the SDs. The sampling system is realized by a winch which lowers a physical probe and 5 samplers up to 50m depth. When the probe and samplers are recovered onboard, the water of the samplers is distributed through the fluidic system to the onboard SDs for water analysis. Finally, the robot is powered with lithium batteries and is equipped with a radio module for wireless communication with the ground control station.
The two hulls contain most of the needed modules. Other modules are located in the deck, a lodging waterproof. The hull is fabricated in carbon fiber and its net weight is about 30 kg. The total weight of the robot, including 4 SDs, is about 90kg and the dimensions are 1.91m length and 1.164m wide. The draft of the boat, including the propellers, is 17.9cm and the waterline is 1.91m. Maximum speed is 3.5 knots; typical speed is about 2.0/2.5 knots.
A dynamic model of the vehicle was produced and an identification session was carried out by using an off-line, sensor-based approach to identify the different model parameters.

The maximum propellers rotation speed is 1400 rpm at which a speed of 3.5 kns is reached.
Fig. 11 shows the surge drag produced by the dynamic model vs the surge speed. At usual cruise speed (1.2 m/s) the robot exhibits a drag of about 40 N.

Two robot catamarans have been completely equipped and validated in the last project year. All the hulls and the deckhouses have been developed using carbon fiber material.
The main components of the robots are:
Supervisor (main controller),
Localization System (Magnetic Compass, GPS),
Radio and Bluetooth Communication Modules,
Locomotion (right/left propeller motor),
Rudders (one for each hulls),
Sampling System (winch/ syringes),
Bio, Optical and Electrochemical SDs,
Obstacle Avoidance SW module using:
- laser scanner, sonar, altimeter,
LiPo Batteries.

A sampler is located at the catamaran stern with a winch that permits to lower a probe.
This probe is composed by:
a commercial probe (YSI 6920V2) with physical and water quality SDs, such as Temperature, Turbidity, pH, Dissolved Oxygen, Oxide Reduction Potential and Conductivity.
five samplers that collect water (each up to 200 ml) at different depths (up to 50 m).
The catamaran communicates in real-time with the ground control station via a 434MHz radio channel to receive navigation and operation commands and to send status and environmental measurements. The catamaran navigates in sea coastal waters with a great stability and precision. The catamaran can steer a planned route avoiding the unplanned obstacles by using proprietary algorithms which integrate data acquired by laser scanner and forward looking sonar. The catamaran shape was designed to permit an easy integration of a sail system to save energy and making the robot visible from long distances (safety). The two hulls contain most of the needed HW modules; the others are lodged in the deck between the two hulls. The sampling subsystem is located at the catamaran stern; a winch permits to lower down till 50m a probe equipped with water physical SDs and syringes to get water samples for the on board-SDs (Hg, Cr, Cd, dissolved hydrocarbons). The SD to detect oil on the water surface is present on an arm at the bow of the catamaran. All the inside devices communicate with the on-board main CPU by a CAN bus; the sampling subsystem communicate via Bluetooth.
All the interfaces (mechanical, electrical, fluidic, communication and software) have been standardised to allow an ease mounting of the current or additional SDs (e.g. Pb).
On March 16, 2012 the catamaran robot #1 performed a long-range mission to Antignano (a village 6 km south of Livorno). The mission aimed at performing a monitoring mission similar to that carried out by Environmental Agencies in terms of covered distance and duration. Furthermore, the mission provided useful information about the robot power consumption and navigation performance in long-range paths.
Table II - The HydroNet catamaran power consumption vs surge speeds.
Surge speed
(m/s) Average power consumption (W)
1.2 230 W
1.3 280 W
1.4 335 W
1.5 370 W
The catamaran navigated autonomously for 12.562km for a total energy consumption of about 820 Wh. Considered the designed battery pack has a total energy capacity of 1800 Wh the test shows the robot can navigate for at least 20 km as required. Furthermore, by looking at the power consumption (230 W) at 1.2 m/s (the cruise speed) it is able to navigate for more than 6 hours (consuming 1380 Wh and covering a distance of about 26 km) with the remaining available energy for the samplings.

Components Description Weight (g) Mission Power Consumption (W) Notes
Main controller Titan PC/104 520MHz PXA270 XScale
Bluetooth module Paranì EST200 to communicate with the Sampling Probe
WiFi module for programming/debugging 650 3.5
Locomotion module Propulsion:
2 × Maxon ECpowermax 30 (200W)
2 × Controller Maxon EPOS70/10
2 × Propellers and actuation mechanism
Rudder motor:Maxon EC 16 (15W)
Controller Maxon EPOS2-50/5
2 × Rudders and actuation mechanism 6120 navigation function Maximum propellers speed 1400 rpm

Maximum rudders speed (20 deg/sec)
Localization module GPS Novatel OEMV-1G-1HZ
Compass TCM 3.0 Pni Corp
NEXUS TH43 paddle wheel water speed sensor
ARM9 STR912F STMicroelectronics
Antenna GPS (external to the hull) 480 3.5
Obstacle avoidance module Laser Scanner Hokuyo UTM30LX + water proof protection
Sonar Tritech Micron DST
Altimeter (GEONAV NAV80003)
Titan PC/104 520MHz PXA270 XScale 1950 16.5 The Titan board collects data from the range sensors, executes the obstacle avoidance/mapping algorithms and communicates through an Ethernet channel with the main controller
PMU Boards with switch regulators 5 and 12 V 200 N/A
Sampling subsystem Winch + motor and frame;
Fluidic system; Sampling probe:
(physical sensors) (YSI6920V2) with
5 PVC water samplers 12500 Active only during the sampling Total sampling probe weight (YSI probe+5 samplers with syringes + control electronics and battery) (8100 g)
Sensor Devices 4 original SDs for measuring concentrations of Hg(II) (or bioavailable Hg), Cr(VI), Cd(II) and dispersed oil;
1 optical sensor for oil slick detection 16000
(est. total weight) 10/each 4-7Kg each
(estimated weight,
except oil slick detection)
Communication module Microcontroller with radio XE1205
Radio antenna (external to the hull) 460 1.5
Cabling & boxes IP67 plastic boxes with waterproof connectors and glands 4150 N.A.
Hulls and deck 29900
Total 72410* 35
*without considering the batteries weight.

Table III - HydroNet catamaran detailed components.

Lessons learnt
A detailed discussion of technical lessons learnt concerning the catamaran has been presented in the revised D4.7 Final Report Floating Robots and Buoys.

It should emphasized the difficulties coming from the very small hull dimensions and the lack of a binding internal layout. Narrow spaces, electrical wires and Teflon have often prevented easy access to the individual components of the catamaran forcing engineers to spend a lot more time in their operations. The internal robot layout should be well considered from the project beginning and strictly observed during all the construction phases. Moreover, an open architecture of the catamaran deckhouse should be applied while maintaining the hosted components safety: even in calm waters, it is really impossible to open the deckhouse to simply have look inside.

Despite of its good precision and accuracy, the localization catamaran subsystem demonstrated poor reliability and robustness: very often the GPS simply do not work and, therefore, the catamaran position is unknown. The flat-boat mounts the same GPS and it works safely. The GPS is not defective, wiring has been redone several times, his position has been raised, but the problem has not been solved. A different wheel water speed sensor should be used, since the actual one do not indicate the rotating direction.

The cabling and communication system also needs some improvements: sometimes communication problem raised over the CAN bus and they was mainly due to the integration of different devices based on different microcontrollers and on the weak robustness of the cabling system. The CAN bus speed should be reduced to 500Kb to improve communication robustness (1Mb communication in this slow vehicle is useless).

The coupling of the sampling subsystem (the probe with the surrounding 5 syringes) with the catamaran fluidic system has revealed weak because a mismatching in the coupling sometimes occurs when the sampling probe is raised into the deck: in fact, during the winding phase of the winch, the bearing mounted on the head of the probe gets stuck on the bearing track of the input column. Moreover, sometimes the syringes do not go correctly into the inlet valves. The jamming was due to many reasons: slight sloping angle of the bearing track, displacement of head and bearing on the probe, imperfect balance of the sapling system. Deformations of the carbon fiber structure of the deck cause the syringes to no go into the inlet valves. The first three points are consequence of the constraints imposed by the limited available space on the deck. The last point is due to low stiffness of the carbon fiber structure. A new design of the winch is already in progress to solve these issues: the new design consists in a new head for the probe directly connected to the rings supporting the syringes. The slope of the bearing track has been increased. Finally, the sampling system tends to drop due to the impact of the hulls with the sea waves.

Also the catamaran deckhouse is very small and closed by carbon walls. This prevents an easy and fast access: the situation worsens as all the logged components need to be demounted and remounted very often in the integration and testing phases. An open architecture should be considered maintaining the hosted components safety (stainless steel use).

Finally, the catamaran should be improved to deal also with worst sea condition (at least sea state 5), it now can navigate in relative almost good sea condition (sea state 3/4).
4.1.3.1.2. Flat-Boat robot
The HydroNet Flat-Boat robot, developed by RoboTech Srl, is a light boat with flat bottom and low draught for navigation in rivers and lagoons for monitoring chemical pollutants (heavy metal and oil) in surface water. The propulsion system of the robot is furnished by two outboard propeller motors placed at the stern of the boat. The control and navigation system (motor drivers and on-board computers) and the navigation sensors (laser, GPS, digital compass and water speed sensor) allow the robot to move autonomously in its working environment through a series of waypoints set by a ground control station. Purposively developed algorithms for navigation (path planning and VFH+) allow the robot to move in its working environment avoiding obstacles.
The robot is equipped of electrochemical and optical SDs (Hg, Cr, Cd and dissolved hydrocarbons), of an outboard SD to detect oil on the water surface and of a sampling and fluidic system (pumps, flowmeter and Teflon tubing) to collect and distribute surface water to the SDs. The robot is powered with batteries and a solar panel to increase energy endurance. Finally, the robot is equipped with a radio module for wireless communication with the ground control station.

The hull is fabricated in carbon fiber and its net weight is about 26 kg. The total weight of the robot, excluding the SDs, is about 75 kg and the dimensions are 2.29m length and 1.12m wide.
The draft of the hull is about 10 cm; the draft of the boat, including the propellers, is 23 cm and the waterline is 1.80m. Maximum speed is 4 knots; typical speed is about 2.5/3 knots.
The propellers have a special design able to cut the seaweeds which are common in river and lagoon.

The motors provide a thrust of 133N, for a total thrust of 266N. The system is powered with 12V, and power is provided by 2 lead acid batteries (12V, 55 Ah each) and by a solar panel (max 19V, 90W) for a total of more than 110 Ah. Power consumption is about 165W at 2.2 knots. Maximum travel distance is about 12.5 km (25Km round trip). Typical endurance is about 6/8 hours.

Demonstrations of the flat-boat and of its navigation capability have been done in Marano Lagunare on June 23, 2011, Most na So?i, Slovenia, on January 12, 2012 and Livorno, Italy, on January 28, 2012. In these demonstrations the robot was commanded by the Ambient Intelligence platform to reach a target point to collect surface water and perform water quality analysis with oil in water SD and Hg bioSD. Additional tests have been carried out in March 2012 on the sea in front of Livorno and in August 2012 in a small lake in Pontedera (Italy).

Application Lagoon and river monitoring of surface water chemical pollutants (heavy metal and oil).
Operational Modes Autonomous navigation with obstacle avoidance capability in river or lagoon through a series of waypoints set by a ground control station.
Station keeping position function opposing to environmental disturbances while sampling the water.
Remote control by an operator from the ground control station.
Navigation Sensors: GPS, digital compass and water current for estimating position;
laser scanner for obstacle detection.
Algorithms: path planning, VFH+ obstacle avoidance based algorithms.
Dimensions L x W x H: 2.29m x 1.12m x 0.58m
Material Carbon fiber
Weight 25kg the hull, ~ 75 kg the robot (sensors excluded)
Hull Draught 25 cm
Actuation Two independent outboard propeller motors, 12V, 30A, thrust 133N
Speed Typical ~ 2.5 Knots, maximum ~ 4 Knots
Power 2 x 12V,55Ah lead acid batteries and 90W solar panel
Power Consumption ~ 240W (12V 20 Ah) at 2.5 knots
Endurance ~ 6 hours at 2.5 knots/~ 8 hours at 2 knots
Maximum distance 12.5 km
Sampling 2 pumps, flow 250 ml/min, surface water only
Teflon piping and flowmeter
Electronic components PC 104 CPU module, Wi-Fi module, Radio communication module; Motors Controller; GPS, Digital Compass, Water current sensor, Laser Scanner
Software Embedded Linux OS

Table IV - HydroNet Flat-Boat detailed components.

In the flat-boat extensive tests executed on March 13th, 2012 in front of the SSSA Marine Robotics Laboratory in Livorno, all measures performed by AE-2 Mini SD presented a low intensity below the value indicating the presence of oil. The measures performed by CRAB SD presented a low intensity below the value indicating the presence of oil slicks almost everywhere except for those taken in the port, which indeed were taken outside of the mission: here the measured values were close to the critical value demonstrating that the SD was able effectively to detect a difference between water outside of the port, usually clean, and inside the port which is typically dirty due to oil discharges from boat engines. During the all day, the flat-boat performed the mission navigating autonomously for 7470m covered in about 2 hours. The average surge speed during the mission was about 1.0 m/s or 2 knots. The overall spent energy was about 325 Wh. Considered the batteries have a total energy capacity of 1320 Wh, this test shows the flat-boat can navigate for at least 30km or 8 hours.
In the simulated long-duration mission, executed on August the 29th, 2012 at Braccinis Lake in Pontedera/Italy, the average power consumption of the flat-boat was 6.6A for the left motor and 5.6A for the right motor, for a total of 12,2A. Considering that the consumption of the other electronics components is about 1.5A the total power consumption was 13.7A that is 165W. Considering that the robot was active for 6 hours the total consumption was 990Wh. Being the batteries capacity equal to 1320Wh and the solar panel provides about 90Wh, we can estimate that the robot had at least 2 hours of energy autonomy remaining. Before the mission the battery was fully charged and its level was 13.2V. At the end of the mission the battery level was 12.2 (Fig. 19).

Fig. 21 show the trajectories of all the 32 rounds. It can noted that robot showed almost the same behaviour and performed almost the same path all along the day, the deviations from the straight lines are not due to the wind effect, but to wrong angles calculated by the digital compass. (Fig. 20). The correction function for the wrong angle has been determined and by applying it angle error is almost close to zero. A more detailed description of obtained data and their analysis is described in HydroNet D3.5 HydroNet Platform Validation in Italy.

Lessons learnt
The flat-boat demonstrated good reliability and fault tolerance: during the different test and demonstration sessions (see D3.9) no particular HW or SW faults and defects came out. Large space inside the flat-boat facilitate the mechanical integration and the cabling of the different components as well as their maintenance

The localization module of the flat-boat, though it would be the same as the catamaran, do not suffer the same problems. Nevertheless it introduces and error in the computation of the robot orientation.
On the contrary, tests showed an high boat resistance to the water possibly due to the non optimal design of the its bow derived from the catamaran shape to save money.

The control loop of the flat-boat navigation system, tuned and tested mainly in calm water (small lakes), needs to be improved: as shown in D3.9 during the sea missions in front of Livorno harbor and in the simulated long-duration mission executed in Pontedera the flat-boat tends to deviate from the planned trajectories and this is due not only to the poor accuracy of the digital compass but also to currents and wind. The navigation system can be improved also adding encoders to the motors.
4.1.3.1.3. Buoys robot
Three different buoys types have been developed in HydroNet by RoboTech Srl: Repeater Buoy, Electrochemical Sensor Buoy, and Atmospheric Sensor Buoy. All buoys are equipped with battery, solar panel, radio module and antenna. The function of Repeater Buoy is to guarantee a total wireless coverage in the robots operating area; Electrochemical Sensor Buoy is equipped with a electrochemical SD; Atmospheric Sensor Buoy is equipped with an anemometer.

The Repeater Buoy is equipped only with the device for the communication infrastructure. Radio module is configured as router dispatching messages between ground station and robots/buoys.
The Electrochemical Sensor Buoy is equipped with on-board computer and sampling system, consisting of a pump, Teflon tubing and waste water tank. The buoy can be equipped with any of the HydroNet Electrochemical SDs. Sampling and water analysis is managed by the control station and AmI software. The Atmospheric Sensor Buoy is equipped also with on-board computer, anemometer and compass. Data from compass (heading, pitch and roll) and from anemometer (wind strength) are integrated by the on-board microcontroller board and computer to compute wind direction and force. Additional atmospheric SDs such as a current meter or a wave sensor can be easily added to the buoy. Sensor analysis is managed by the control station and AmI software.

Application Routing system to increase wireless coverage of robot operating area
Dimensions D x H: 0.7m x 1.10m (1.35 including the antenna)
Material Polyethylene and fibreglass
Weight ~ 38 kg
Power 1 x 12V,7.5Ah lead acid batteries and 10W solar panel
Power Consumption ~ 0.05 Ah at 12 V
Endurance infinite lifetime
Electronic components Radio communication module and antenna
Table V - HydroNet Repeater Buoy specifications.

Application Analysis of heavy metals (Hg, Cr, Cd) and oil in water
Dimensions D x H: 0.7m x 1m (1.25 including the antenna)
Material Polyethylene and fibreglass
Weight ~ 55 kg
Power 1 x 12V,7.5Ah lead acid batteries and 10W solar panel
Power Consumption ~ 1 Ah at 12 V
Endurance > 7.5 hours
Sampling 1 pumps, flow 250 ml/min, depth up to 50 m depending on pipe length
Electronic components Radio communication module and antenna, PC 104 CPU module
Sensors Devices Hg, Cr, Cd, dispersed oil
Software Embedded Linux OS
Table VI - HydroNet Electrochemical Sensor Buoy specifications.

Application Acquisition of atmospheric signals (wind)
Dimensions D x H: 0.7m x 1.10m (1.90 including anemometer)
Material Polyethylene and fibreglass
Weight ~ 47 kg
Power 1 x 12V,7.5Ah lead acid batteries and 10W solar panel
Power Consumption ~ 0.5 Ah at 12 V
Endurance > 15 hours
Electronic components Radio communication module and antenna, PC 104 CPU module
Sensors Devices Anemometer, compass (3-Axis tilt-compensated)
Software Embedded Linux OS
Table VII - HydroNet Atmospheric Sensor Buoy specifications.

Lessons learnt
The buoy improvements basically concern their dimensions which now results too small and consequently are almost unstable in the water: for simplicity and logistic reasons small and light buoys with limited buoyancy were used in the HydroNet project. Small buoys are easy to transport and deploy in the HydroNet demonstration sites without the specialized operators help. As reported in D7.8 higher performance buoys are needed in a real deployment such as the SLB1250 Navigation Buoy by SeaLite (http://www.sealite.com.au). The SLB1250 buoy is more robust and strong with respect to the current HydroNet buoys, it has a diameter of 1250mm, a height of 1180mm and the mass is 70 Kg. It is provided with all the devices and features (color, lights, signals. etc) needed for immediate deployment in sea. This buoy is also already provided with an internal box for housing electronic components and SDs.
4.1.3.2. HydroNet Sensor Devices (SDs)
The core of the HydroNet platform is represented by biological, optical and electrochemical SDs with communication capabilities.
New miniaturised, original, custom-made SDs (20x20x30cm, 5-7 kg each) have been developed by the HydroNet Consortium to detect several pollutants (chromium, cadmium, mercury, hydrocarbons, oil) in water bodies. All the different SDs (electrochemical SDs for Cd(II), Cr(VI), Hg(II), bioavailable Hg SD, optical sensor for dissolved hydrocarbon and surface oil) can be lodged into the robots through standard electrical and fluidic interfaces (see Fig. 15). This guarantees the possibility to easily insert into the available slots the different types of available SD before a mission. The Hg bioSD lodged in one robot locker is showed in Fig. 15. The HydroNet SDs characteristics vs the EU limits values are shown in Table VIII.

Chemical species Law limits HydroNet
results HydroNet types
Cadmium (II) 1 µg/l 0.05 µg/l HUJI electrochemical
Chromium (VI) 50 µg/l 1 µg/l HUJI electrochemical
Mercury (II) 1 µg/l 1 µg/l HUJI electrochemical
Mercury bioavailable 1 µg/l 0.01 µg/l IFB bio-sensor
Total Petroleum Hydrocarbons (dissolved in water) 0.2 mg/l 0.05 mg/l LUMEX optical
Oil slicks on water surface Yes/No Yes/No LUMEX optical
Table VIII - HydroNet Sensor Devices Characteristics.

Lessons learnt and conclusion
During the firmware development, all the SDs teams met the challenge of the interface realization for CAN protocol that is commonly used by the industry for marine and terrestrial vehicles and that is rarely associated with SDs and analytical instruments. This required the inclusion in the original chemists/biologists group of persons with computer skills. Moreover, the need to upload/download water to clean the internal tubes/cells and to measure the required element/substance prescribed the use of pumps and valves and required the adoption of mechanical and hydraulic expertise. The end result has been the creation of a multidisciplinary group in the SDs developers laboratories and exciting feelings in the working place.

The developed SDs need to be optimized from an electronic/electric point of view. The maintenance operations of SDs has to be made easier. The electrochemical SDs and the bioSD need to be fully validated and certified to be accepted by the environmental agencies for real monitoring campaigns,

All the developed HydroNet SDs meet the revised requirements in the middle term review of the project (July 2010) and are able to detect the EU law limits (some SDs are performing even better). All SDs are enough are light and small and almost meet the constraints imposed by HydroNet floating boats. The electrochemical SDs are better of any other similar (same dimension and weight) device available on the market.
A new electrochemical cell to analyze a water flow has been designed and realized by the HUJI partners and a patent application is in progress.
4.1.3.2.1. Electrooptical SDs
All the HydroNet optical SDs were developed by Lumex, a Russian SME.

HydroNet oil in water optical SD
The Fluorat AE-2 Mini oil in-water SD is a compact and automatic miniaturised sensor for measuring petroleum hydrocarbons in water. The oil-in-water SD uses fluorimetric method of registration of total petroleum hydrocarbons (TPH) based on original membrane-extraction technology. TPH are extracted to membrane from water flow using hexane. Fluorescence intensity of the TPH extracted to the membrane depends on extracted THP quantity. The UV LED is used as excitation light source and photodiode as a detector.
A value of registered fluorescence intensity is compared with tabulated figures obtained during the sensor calibration with TPH standard reference materials (SRM) and thus the value of the TPH concentration is calculated. The concentration value is transferred to central processor using CAN protocol. The calibration coefficient is stable, and the sensor should be recalibrated using the SRM every 30 40 days. The SD box is provided with power cable, CAN bus connection and input and output connections for water. External view of the new SD is shown on the Fig. 27(left), internals in the Fig. 27(right). Technical specification of the SD is listed Table IX.

The main strengths of the Fluorat AE-2 Mini SD are:
- Selectivity: all hydrocarbons extracted by hexane are considered as TPH.
- Full TPH extraction from a sample due to the sample circulation inside extraction module.
- Block structure of the sensor: easy diagnostics and repair by blocks replacement.

Further work should be focused on possible improvement of optical and hydraulic schemes to decrease limit of detection and minimise sample volume and hexane consumption. For example, existing optical scheme could be improved by use of more powerful light sources (LEDs).
Application Oil in water concentration (mg/l)
Sensitivity Min 0.5 mg/l, max 10 mg/l
Dimension L x W x H: 200 mm x 200 mm x 300 mm
Weight Min 7 Kg , Max 8 Kg
Power 12 V
Power Consumption Sleeping mode: 1 W
Average: [7] W (2 measurements per 1 hour)
Max peak power in Operating Mode: 32 W
Working Temperature +5, +400C
Water needed to measure + rinse Min 200 + 200 ml / cycle
Calibration Every 30 days. Single needed time: [3600] sec
Off-Shore Recondition Times Substituting consumables: 3600 sec (if any)
Sampling Time 300 Sec (measurement only), 900 Sec (with internal sensor cleaning)
Discharged whole liquid quantity 200 ml / cycle
Discharge mode Tube n° 2 (dependent mode) or Tube n° 3 (autonomous mode)
Life Life-measurements: [100]
Life-time: 3 months
Returned Values Output messages in the same measurement: 1
Concentration: value (signed integer)
Returned Values Resolution: [0.001]
Cable Power cable: + 12 V > brown; ground > blue
Signal cable CANH > red; CANL > yellow; ground > black

Table IX - HydroNet Fluorat AE-2 Mini SD specifications.

HydroNet oil on water optical SD
The HydroNet oil on water optical SD, CRAB, is designed for continuous automatic oil slicks detection on water surface. The principle of operation is remote optical sensing of water surface with laser beam. Oil slick detection is based on difference of light reflection from clean water surface and oil slick.
Detector is comprised of a light source (laser diode), one-channel optoelectronic sensor and processing unit. The intensity of the laser radiation is modulated on frequency. Optoelectronic sensor is comprised of optical diode, preliminary and selective amplifiers and detector. Statistical processing of signals is done by the processing module using embedded microprocessor ATMega 16 within the prescribed time (30 sec). The measurement result is sent to the central processor through CAN interface. Above clean surface, normalized signal doesnt exceed 1 V. If the result exceeds the threshold value (1.2 V) the alarm signal is generated. The minimal detected thickness of oil film is estimated as 0.5 micron.
The main strengths of the CRAB SD are:
- Remote sensing: no sampling, continuous water surface monitoring,
- High sensitivity: detection of oil slicks with oil film thickness > 0.5 micron,
- Low interference from ripples: no affect from optical axe/surface angle up to ± 20 grad,
- Low interference from scattered radiation due to light modulation,
- Built-in processor for data acquisition and processing, non-volatile memory,
- Robust performance for outdoor installation (IP66 housing).

Further work should be focused on possible improvement of optical scheme to expand installation height above surface (now it is 0.7 1.2 m, adjustable up to 15 -20 m). This will expand application area: installation on bridges, off-shore platforms, etc. Finally, different systems of data transfer could be developed depending on application task.

Application Parameters detect: oil slick detection, zero surface slope probability.
Remote sensing of water surface with a light beam based on difference of the probing light reflection from water and oil slicks.
Sensitivity Min 0.5 micron
Recommended height 0.7 1.2 m
Dimension L x W x H: 150 mm x 150 mm x 100 mm
Weight 1 Kg
Power 12 V
Power Consumption Sleeping mode: [0.3] W
Average: 1.2 W (continuous measurement)
Max peak power in Operating Mode: 1.2 W
Working Temperature -10 - +50 0C
Water needed Not necessary
Calibration Not necessary
Off-Shore Recondition Times Not necessary
Sampling Time 30 Sec (measurement only), no internal sensor cleaning is required
Discharged whole liquid quantity No discharge
Discharge mode No discharge
Life Life-measurements: 6 months [nn]
Life-time: [6] months
Returned Values Output in the same measurement:
1. Zero surface slope probability, value (signed integer)
2. Value (two bytes, lower byte first): mean square of radius of surface curvature (arbitrary units), signed integer, range min 0, max 5000
3. Reflection coefficient, value (signed integer)
Value (next two bytes, lower byte first): reflection coefficient (arbitrary units), signed integer, range: min 0, max 5000.
Cable Power cable: + 12 V > brown; ground > blue
Signal cable Signal cable: CANH > red; CANL > yellow; ground > black

Table X - HydroNet CRAB SD specifications.
4.1.3.2.2. Electrochemical SDs
The HUJI developed electrochemical SDs for three heavy metals: cadmium, mercury and chromium. The SDs were assembled in the robot and are able to carry out autonomous calibration, measurement and transmitting the results to the central control station. Since the SDs box were to be attached to a sample supply system that should provide a sample of water to be analyzed, it was clear that an internal flow system needed also to be developed. Indeed, such flow system was successfully developed (see below) and the electrochemical cell was integrated into this flow system. Hence, two major activities have been carried out: construction of the flow system, which is more a technological/engineering task, and the development of three electrochemical SDs that will first work in a static system and then to incorporate them into the flow system. These two goals were completely achieved. Furthermore, the requested sensitivity of the electrochemical SD was drastically reduced after approximately one year of the beginning of the project. This decision was made as a result of the request by the Jozef Stefan Institute / Slovenia partner such a way to be able to use the SDs for non-polluted systems. Hence, the requested sensitivity was reduced by a factor between 20 and 100, which made the chemical goal significantly more challenging.
All in all, the output of this project is impressive. A huge advancement has been made and the transduction part has been successfully coupled with the flow system. A few versions were constructed and a unique electrochemical cell was designed that was patented. On the other hand, the SDs validation has not yet fully completed, a mission that will be soon addressed. National funds for this tasks have already been raised and a MSc student was recruited for this mission.

The electrochemical SDs characteristics (sensitivity, range of linearity, weigh, volume etc.) are shown in Table XI.

Hg(II) Cd(II) Cr(VI)
5 5 5 Power Consumption (W)
12 12 12 Power Supply (V)
3 3 3 Weight ( Kg )
10 mL 10 mL 10 mL Sampling Water ( l )
30 min 30 min 30 min Measurement time
1 week 1 week 1 week Sensor maintenance
< 1µg/L < 0.05 µg/L < 1µg/L Detection limit
± 20 % ± 20 % ± 20 % Accuracy
1-100µg/L 0.05-10µg/L 1-100µg/L Range
50 mL 50 mL 50 mL Volume
RS232/CAN BUS/Analogic signal Output Signal
yes yes yes Oriented (yes/no)
10-50 mL 10-50 mL 10-50 mL Waste production (for measurement)
every day every day every day Calibration Period (Manual)

Table XI - HydroNet electrochemical SDs characteristics.

The HydroNet obtained results, described in the following, are divided to the technological part that deals with assembling the flow system and integrating the SDs core and the electrochemical part, which aims at constructing three different electrochemical SDs core.
Flow system
The first prototype of our flow system was prepared using the assistance of the project partners from Slovenia (IFB). The first version of our system, which can be seen in Fig. 29a was introduced in the Livorno mid-term review. A scheme of the system is shown in Fig. 29b.
Since then, we have had major progress in three paths: the electrochemistry, the electronics and the engineering of the system. From the electrochemistry point of view we had to solve a few fundamental problems arising from the changes in the dimension of the electrochemical cell and the arrangement of the electrodes. The first problem we encountered was emerged from the electrical field. In Fig. 30a there is a drawing of the first electrochemical cell we examined. The problem was associated with the configuration of the electrodes in the cell, i.e. the working electrode was not aligned and located far from the counter electrode. From the electrical point of view, we had to manufacture an automated, computer-controlled sensor. We developed with the help of our electronic support team, a LabView based program, which controls our entire system. This system utilized a PalmSens potentiostat (Emstat) and a Eurotech microprocessor (Cd SD) and an improved microprocessor which was manufactured by Avalue, Taiwan (for the Hg and Cr SDs).

Fig. 29 - Picture of the early versions of the flow system. (b) Schematic drawing of the flow system.

This caused interference in the electric field. In order to overcome this dilemma, we designed and manufactured the cell shown in Fig. 30b. In this cell, both electrodes are aligned, and the electric field that is generated is symmetric. Fig. 30c shows schematically the combined counter and reference electrode that was fabricated for cell Fig. 30b. The combined electrode suffered from bad sealing, this was the motivation for the final design showed in Fig. 30d. In this cell we still have a perfect alignment between the working and the counter electrode, and we also solved the problem with the reference electrode. This flow-cell proved to be very efficient and in all of our flow systems, the electrochemical cell, which is the heart of our SD, is based on the cell described in Fig. 30d.

HUJI has already filed a patent application for this cell configuration (U.S. Application Number 61/663,750), which is unique and does not exist commercially. All of our SDs uses a similar program, which differs only by the electrochemical scan method (square wave or linear sweep voltammetry). This concept simplified the SDs integration. In fact, the integration problems that arose with our 1st sensor (Cd), made the integration of our 2nd sensor (Cr) an easy task. From the engineering point of view, we had to manufacture a sensor that needs to store liquids as well as electronic and electrical devices in very small dimensions. Using the help of out mechanical workshop crew, we made numerous adjustments and modifications in our first sensor (Cd), which can be seen in Fig. 30b, until we were able to construct a sensor that not only meets the dimensions of the HydroNet project but also the weight requirements. In Fig. 31a there is a picture of our improved sensor design. Two major changes have been made; the electronic equipment was moved to the top of the sensor making the handling simpler and also reduces the risk of accidental damage to the electronic circuits. The second improvement was the arrangement of all the pumps and liquids in a way that not only reduces the risk to the pumps from mechanical damage or from liquid spills, but also makes the sensor internal less voluminous and easy to handle.

Cadmium electrochemical sensor
The sensor for Cd was the first and has been very successful in operating it in the lab. We published a paper in a high impact factor journal, which describes our results. The sensitivity of this sensor reached the new level that is 50 ppt (parts per trillion). We were able to measure the levels of Cd in drinking water and compared that with a standard measurement by ICP-MS. The sensor is based on a self-assembled monomer, i.e. of a monomolecular layer that assisted in the deposition of the low concentrations of the metal on the electrode.
The sensor utilizes a novel approach where very low levels of Cd(II) can be determined electrochemically based on the under potential deposition (UPD) of cadmium on a self-assembled monolayer (SAM) modified Au electrode. In order the reach the level of detection (LOD) desirable for the HydroNet project we used a pulse method (square wave voltammetry) combined with a subtractive method. The initial level of detection for Cd sensor, as set by the project objectives was 1 ppb, later this objective was altered and a new goal of 0.05 ppb was set. Our sensor was capable of detecting 0.05 ppb of Cd2+. In the self validation experiment we conducted, we were able to measure the levels of Cd in drinking water (between 0.1-0.3 ppb) reproducibly. The results where compared to the standard method (ICP) and the results where almost identical. The results were published in detail in Analytica Chimica Acta 2010, 2010, 684, 1-7. Furthermore a detailed manual for this sensor was also written.

Chromium electrochemical sensor
The sensor for Cr(VI) is based also on a self-assembled monolayer possessing a pyridinium moiety. The electrode is also made of gold and the level of detection reached that of the new level, which is in the ppb range. Yet, the electrode requires further optimization and validation.
The sensor for chromate, i.e. Cr(VI) is based on the same gold electrode used for the Cd sensor, however, modified with a different thiol self-assembled monolayer. Cr(VI) which exists in water as CrO42- (neutral to slightly acidic pH) is selectively extracted by a pyridinium based monolayer that is attached onto the gold electrode by adsorption. The electrode is made by first polishing and electrocycling it in a clean acidic solution, followed by immersion in the thiol solution for less than an hour to form a stable monolayer. The electrochemical detection is carried out by concentrating the chromate by the modified electrode under either open circuit potential or by applying a constant potential for a few minutes depending on the concentration of the Cr(VI) in the solution. For lower concentrations a longer time of extraction is needed, however, this period does not exceed 15 minutes. Then, the potential of the electrode is scanned from positive to negative potentials, which causes the reduction of Cr(VI) to Cr(III). This also releases most of the chromium from the electrode surface and makes it possible to reuse the electrode numerous times. We fabricated two electrode prototypes one based on 4-(2-mercaptoethyl)pyridine that was synthesized by us, and the other on 4-mercaptopyridine, which is a commercial product. The two electrodes showed good reproducibility and we have been able to measure Cr(VI) using our automated flow system. Although the project reached its end, we are still working on a modified version of the electrode which is planned to replace our current electrode in order to improve out LOD. We already synthesized a unique derivative of pyridinium polymer and will explore its ability to extract more efficiently Cr(VI).

Mercury electrochemical sensor
The sensor for Hg was of particular interest to the other partners. Two versions were developed one that is based on a glassy carbon electrode, which gave high sensitivity but not high reproducibility, while the other takes advantage of a gold electrode. In both cases the level of detection was below 1 ppb but we have not reached the new level of 10 ppt (we are pretty close to it). This electrode still requires further validation and testing in the flow system. Two more papers originated from this study. One was already published while the other is under preparation.
In the course of preparation of the Hg sensor we had to take into consideration different restrictions and limitations, such as short measurement time, high reproducibility, small volume, environmental issues and affinity to mercury as well as selectivity. These conditions derive both from the electrochemical process and the mercury itself. The working electrode selection is one of the most important parameters in electrochemical detection.
The first working electrode chosen for mercury detection was made of glassy carbon (GC) which is electrochemically inert over a wide range of potentials, isotropic, very hard and stable in many corrosive media, has low porosity and can form strong covalent bonds with several modifiers. The electrochemical method was chosen to be subtractive Linear Sweep Voltammetry (LSV). Using the GC electrode, which we also modified by three different approaches, we achieved good results that were published in Journal of Electroanalytical Chemistry 2010, 649, 153-158. Unfortunately, the LOD of the GC electrode was not sufficient and we had to seek new directions. In the last couple of months we have developed a new electrode that is based on a selective layer attached onto a gold surface. For a long time we tried to avoid the use of a gold based electrode, since mercury tends to form irreversibly an amalgam with Au. Our experiments show that in concentrations below 100 ppb we can regenerate out electrode using a simple electrochemical process that is compatible for a flow system. Furthermore using self assembled monolayers and our experience not only that we were able to improve our signal and LOD, but also prevent the fouling of the electrode and the formation of an amalgam. We have modified the gold electrode with different monolayers, such as 1,5-pentanedithiol, sodium-3-mercapto-1-propanesulfonate and 3-mercaptopropaneacetic acid. The electrodes show great promise and we were able to reach a LOD of ca. 1 ppb.
4.1.3.2.3. Biological SD
A standalone on-line bioSD for reactive mercury using fluorescent and luminescent bacterial transformants was developed for use in complex aquatic environments. The devices measure bioavailable Hg and general water toxicity. The later defines the presence of toxic substances, which needed to be carefully examined afterwards. In order to sense bioavailable Hg the fusion of biological part with the mechanical and electrical solutions was performed. Microorganisms are the sensing part of the whole bioSD. Moreover, we chose microorganisms, which are the most relevant for the particular environment. For salt water the organisms of choice are Vibrio fischeri and for freshwater Pseudomonas putida. Both these organisms are frequently present in each environment. The signal, which is bioluminescence, of the sensing bacteria, is then transduced to the photomultiplier, which at the end converts light signal into the electronic signal, namely volts and counts. According to the standard curve, Hg concentrations are correlated to the output signal on the photomultiplier. Consequently, the Hg bioavailabe concentrations are then determined for particular sample.
The HydroNet Hg bioSD is now an automated sensor that measures Hg bioavailability and general toxicity of water samples.

Sensor can detect: Bioavailable Hg fraction in water samples and general toxicity of water
No. of SDs delivered: 4
Sensitivity: 10 ng/l (laboratory environment)
100 ng/l (working environment)
Working Range: 10 - 10000 ng/l
Working Temperature Range: 10 - 30 °C
Sample size: 50 ml
Operating voltage: 12 V
Power consumption: 3,2 W (on average)
1,8 W (basal consumption)
Weight: 6 kg (7 kg including liquids)
Running cost / Year: 1000
Full sensor cost: 8000

Table XII - HydroNet bioavailable Hg SD characteristics.

MSDS Specifications and hazards
Bacterial strains used in the Hg BioSD are not pathogenic and harmful to humans. Nevertheless, good laboratory practice (GLP) should always be practiced when dealing with bacteria in laboratory environment, where bacterial cell count is high. Revival of stored bacterial cultures, preparation of bacterial cells for transfer into the chemostat and the transfer of cells into the bioreactors are crucial steps, where GLP should be implemented. In any case bacterial cultures must not be ingested, spilled or otherwise released into the environment.
Bacterial strains used in the Hg BioSD are genetically modified organisms (GMOs), have been assessed as GMO safety levels 1 (Pseudomonas putida) and 2 (Vibrio fisheri), are not harmful to humans and do not a pose and environmental safety hazard. Nevertheless, they should be regarded as GMO material and all safety precautions should be taken when dealing with this material. Discarded cells should be inactivated by autoclaving, waste materials and liquids that have come in contact with the cells should also be autoclaved before being discarded as normal waste.

Schematic representation of the bioSD
The Hg BioSD is composed of several components (the chemostat, the measuring module, the microfluidic system and the air pumps), which are controlled by the microcontroller. The microcontroller ensures that all processes are fully automated, so that the bacterial cells are being kept at the right developmental stage appropriate for measuring. The microfluidic system ensures the cells and the water sample are transferred to the measurement module and that all the internal compartments stay clean and are not contaminated by bacteria living in the environment.

BioSD Measurements
Bioavailable Hg. Bioavailable Hg concentration is calculated from the physiological response of bacterial cells to the present bioavailable Hg. For the determination bioavailable Hg concentration cells with responsive genetic elements are used, meaning the genetic element is expressed only in the presence of Hg2+ ions. The expression of this genetic element finally leads to light emission (luminescence). The concentration of bioavailable Hg translates to light intensity.
General water toxicity. General water toxicity can be predicted from inhibition of the light signal when the performance of bacterial cells (the luminescence) is attenuated due to toxic compounds in the water sample. For the determination of inhibition bacterial cells with constitutively expressed genetic elements are used. In clean samples the constitutively expressed genetic elements cause maximal luminescence possible, whereas this signal is lowered when substances that lower cell performance are present in the water sample. Inhibition is expressed as the percentage of signal attenuation and gives an indication of water toxicity of a given sample.
4.1.3.3. HydroNet environmental pilot sites modeling
Scenarios definition and platform validation
The performance evaluation of the integrated HydroNet system was carried in three selected case study areas: (i) So?a/Isonzo River and the Gulf of Trieste, (ii) Marano Lagoon (Italy) and (iii) coastal area of Livorno (Italy). The selected case studies offer a variety of environments and conditions under which HydroNet platform was tested and validated: rivers, artificial lakes and lagoons, and coastal waters including the mouth of rivers and/or proximity of industrial drainpipes.
These pilot sites were selected based on the previous knowledge on behavior (sources, fluxes and concentration levels) of different pollutants (multiagents) in different scenarios under different environmental conditions/events.
Focus was on possible use of robotic autonomous agents, especially in environments inaccessible to human operators or in situations where the robots can provide effective methods to explore/monitor large areas without or with a limited human intervention. An overview of environmental scenarios under which robots are expected to operate is given in Fig. 34.
The WP3 of the HydroNet Project focuses on the use of existing and new data to characterize sources of heavy metals; the concentrations of heavy metals in water, sediments, soils, biota, humans and wildlife and; the rates of key heavy metals transformations that affect bioavailability; and, the compilation and documentation of all data that were produced before or during the course of the project.
In order to evaluate the performance of the integrated HydroNet system with changing environmental and meteorological conditions, the project included field and laboratory measurement campaigns carried out during different seasons in selected case study areas. Information on inventories published or reported so far for selected sites was provided. Database of metals and its species environmental concentrations and inputs was crucial for modeling purposes and planning of sampling and measurements campaigns performed within the project to provide new data on sources, fluxes and concentration levels of selected pollutants in coastal area of Livorno-Piombino, Marano and Grado lagoons, and the Gulf of Trieste with So?a/Isonzo river Fig. 35.
Missing information about Hg speciation and other metals in water column, sediments and biota were collected at selected sites during several sampling campaigns of sediments and water. Those information were crucial for further modeling work and validation of HydroNet system.

Fig. 35 - Sampling location within HydroNet project in Livorno coastal area, Marano-Grado lagoons, and the Gulf of Trieste with So?a/Isonzo river system.

The set of pilot sites identified for the HydroNet Platform gives the opportunity to test and demonstrate the systems in different environmental settings (as well as in different cultural contexts, with different attitudes towards technology and robotics). The test also aims at demonstrating the functionality and potentiality of the HydroNet platform and to evaluate the performance of the system from a user and technological point of view. Both, river and marine robots as well as buoys were tested and demonstrated in freshwater river So?a/Isonzo, at artificial lake at Most na So?i in beginning of January 2012. Event took place at two locations. Field presentation of robot SDs was performed at artificial lake at Most na So?i, while Gosti?e Rutar in Tolmin hosted theoretical presentation of the project to invited reporters and representatives of local community, governmental and private institutions and agencies, as well as people from Universities and research institutions.
The performance of newly developed SDs were tested and compared with classically validated methodologies to demonstrate the functionality and performance. It was done in laboratory and in teh field (Most na So?i, Livorno coastal area).

Transport-dispersion models
The development of an integrated model (Fig. 36), upgrade of individual models and modules and development of the necessary interfaces (arrows) for integration in-between the models represent significant improvements in dispersion modeling techniques in coastal regions. The models in red were either developed or significantly improved in the framework of the HydroNet project. Circulation in the domains under study is computed using relatively fine grid and all important forcing factors. Adequate boundary/initial conditions for hydrodynamic simulations are acquired from other meteorological and circulation models.
The most vital parts of the integrated modeling tool (Fig. 36) are the two transport-dispersion models (green background), the PCFLOW3D-TD and the NAFTA3D models. They compute transport, dispersion and fate of dissolved and particulate pollutants, and the oil-spills based on circulation computed by either PCFLOW3D HD or any other hydrodynamic model included. River inputs are provided by a GIS-based catchment model. Sediment-transport and sediment-water flux modules are also included into the PCFLOW3D-TD model. The integrated model is supported with measurements (orange) and tools for graphical representation of the results (magenta background).
The model was calibrated and validated against measurements in the Gulf of Trieste (hydrodynamics, sediment transport, diffusive fluxes) and by a comparison to other similar models.

Fig. 36 - The HydroNet integrated modeling tool modules (rectangles) and interfaces (arrows).

The modeling tool was successfully integrated into the AmI platform; both measured data and modeling results can be exchanged between the model and the AmI and further used in both applications.

Lessons learnt & conclusions
Deliberations with dispersion models users at various institutes and regional environmental agencies in Italy and Slovenia have convinced us that a fully, tight integrated dispersion model as it should be developed in the HydroNet project is not necessarily an advantage. The end-users usually prefer either to continue developing their own models or using commercial models they are used to work with. Therefore, the PCFLOW3D dispersion model and its sub-models were only loosely coupled into the AmI. In this way, potential end-users of the HydroNet platform (robots, SDs, buoys and AmI) can decide whether to couple their own software into the integrated system by developing simple interfaces or to use the dispersion models developed and upgraded within the HydroNet project.
Coupling a 2D Finite Element Model (shyfem, ARPA FVG) with a 3D Finite Volume model (PCFLOW3D) is a very demanding task. Development of adequate interfaces took much longer than expected and was not finished by the end of the HydroNet project. However, the collaboration established between the model-developing teams (UOL and ARPA-FVG) remains very strong even after the project: researchers at both institutions have agreed on further mutual support and cooperation on the development, coupling and use of circulation and dispersion models in the Grado and Marano Lagoons and the Gulf of Trieste.
Upgrading the models and modelling procedures from seasonal to nearly real-time modelling was a very difficult task. Converting data from the relatively low-resolution numerical grids used in other projects (e.g. MyOcean, approx. 5x5 km) to the UOL grid (500 x 500 m) in the Livorno coastal area was a process of trial and error. Moreover, some interfaces had to be rewritten during the project, as the format of the data provided by the supporting projects had changed. Much better results were achieved in the Gulf of Trieste, where simulations in an intermediate resolution (approx. 900 m) were used as the open boundary condition. Sequential nesting of modelling domains in this area finally resulted in performing test-simulations using a very fine grid (40x40 m) in shallow coastal parts (Koper Bay).
4.1.3.4. HydroNet Ambient Intelligence infrastructure
The floating, sensorised, autonomous robots are part of an Ambient Intelligence (AmI) Framework that encompasses all the software infrastructure that is responsible of interconnecting the HydroNet fleet of robots, human operators and intelligent software modules in order to manage and task the fleet, monitor real-time all the events that occur at a HydroNet site and produce value adding information about the observed environment. The AmI Framework acts as a service level platform interconnecting all the actors (robots and buoys, operators, users, internal or external software services) of a HydroNet site. The AmI Core, part of the AmI Framework, encompasses all the services deployed for storing and providing access to the information produced by the HydroNet robots fleet of and includes also all the web interfaces deployed for configuring, tasking and monitoring. The web interfaces are usually referred to as AmI Clients. The AmI Server contains all the software needed to run the AmI Core.
The abstract robots architecture is been based on a well-assessed three-layer architecture: Planning Layer, Executive Layer and Behavioural Layer. The AmI Framework includes the Planning Layer of the HydroNet robots architecture. To implement the topmost architecture layer an intuitive system to support the definition of tasks to be executed by the robots has been developed: a close-to-natural language approach is used to express commands, timing and synchronization constructs.
In order to be able to handle the large amount of different interfaces, protocols and interactions in a lightweight and loosely coupled way, the AmI Framework includes a Dedalus tool developed for providing Service Oriented functionality and adapted within the HydroNet project to become a minimal Enterprise Service Bus (ESB). This guaranties AmI Core extendibility during the HydroNet project and future developments. On top of this service oriented platform two kinds of data storage services have been implemented for optimizing both real-time, dynamic, flexible accessibility and bulk, resource hungry data transferability.
Several user interfaces have been developed exploiting the AmI Core as service provider. In particular, the AmI Core integrates, through a common browser, a set of rich Internet applications.
Innovative interface for mission planning, based on the Lexicon widget
Lexicon is a very intuitive tool that leads an operator through the composition of a natural language sentence that represents the command to be transparently translated into computation language. Complete missions can be composed of an arbitrary set of commands chained together (in a graph-like structure) through synchronization events (parallel commands , on-failure conditions/commands, on-success conditions/commands). There are special dialog windows that enhance the user experience when dealing with commands related to navigation (e.g. map-based dialog for selecting points and paths) and configuration (robot outline with functional areas).

Easy-to-use and intuitive interface for configuring the robot missions
All aspects of a robot configuration have been summarized in a single interface that shows the robot outline subdivided into functional areas. By clicking on each area the user can open a configuration dialog window for reading or setting the configuration parameters related to the area. In particular it's possible to visualize the SDs allocation on the boat exploiting a Plug aNd Play notification sent by the robot at start-up and to assign them to a syringe configuration by using the SDs annotation, as well as to establish the sampling depth and the amount of water to be sampled.

A content rich interface for monitoring a HydroNet fleet
As main requirements for the monitoring interface it was established that all the HydroNet fleet elements (both robots and buoys) should clearly be visible in both a georeferenced way (robot moving on a map according to its real position) and a logical way (location independent list). Through a very limited amount of gestures it is possible to access detailed status information, configuration and tasking dialogs for every single fleet component. An always visible log console showing real-time all the HydroNet Control Language notifications coming from the fleet is available.
Within the AmI Core, a functionality that produces Keyhole Markup Language (KML) data for external services (e. g. Google Earth), has been developed. This service allows to render in a very appealing and comfortable 3D representation the whole information regarding a HydroNet fleet.
Moreover ATOM has been used as a HTTP based protocol for exporting information to the external world of software producers but also Cloud services and mobile applications. Exploiting this protocol as a glue, a complex interaction for downloading and uploading data between the AmI Core and the PCFLOW3D software from UOL has been designed, implemented and tested.
Concerning the main HydroNet goal to introduce computational intelligence in the work-flow of a typical application scenario, with the goal to support the definition and the application of knowledge management processes for decision support, an intelligence automation service, called Algo Service, has been developed and connected through a very flexible HTTP REpresentational State Transfer (REST) based Application Programming Interface (API) to the AmI Core. In this way, it has been possible to automate tasks like recognizing and presenting critical situations on observational data and to transform textual archives uploaded by PCFLOW3D operators into visual layers to be rendered in every geo-referenced SW supporting KML.
The most important algorithm integrated into Algo Service is the algorithm for pollution source finding: EPSO, a particle swarm optimization algorithm for pollution source finding. Starting from a theoretical implementation of EPSO, in the HydroNet project an improved version of the algorithm has been developed, by including calculations related to:
irregular coastlines and shoreline detection,
pollutant distribution that varies in space and time,
water current and depths variations of the sea.

All the information needed to set up such a complex scenario has been provided by PCFLOW3D data and complex interpolation procedures have been realized to get a very good approximation of the reality and its evolutions.
For implementing the Executive Layer a breaking-through choice has been made by using the LUA dynamic scripting language. This choice proved to be successful in terms of flexibility, maintainability and ease of deployment, which are the key issues in developing robotic prototypes in high variability scenarios.

Moreover, a simple serialization protocol (HydroNet Control Language - HCL), based on reflexive LUA tables, could be developed in order to implement the exchange of missions and feedback notifications between the Executive Layer and the Planning Layer. At the project end of a complete specification of the protocols from Planning Layer to Behavioural Layer has been released consisting of 11 fully documented commands with 40 different facets (Navigation, Debugging, Probe based sensing, Real-time sensing, Configuration, Operating System). At the Behavioural Layer, two modules that implement the Radio Frequency communication and a robot wide configuration database have been developed; simulators of the modules for navigation, system feedback and sensing have been implemented in order to test the overall architecture. Missions and feedbacks are exchanged between the Executive Layer (exploiting the Communication Behaviour on the robot) and the Planning Layer through the Base Station Gateway as compressed messages in order to save bandwidth and exploiting the streaming functionality provided by the radio module SW for large bulk transfers.

Several tests performed in bulk procedures and in the close-to-reality simulated scenarios (Mar Ligure and Gulf of Trieste) showed impressive results. The implemented algorithms have been able to guide a HydroNet fleet towards the pollution source in complete autonomy during missions that lasted for several hours. The testing scenarios of Mar Ligure, Isonzo and Piran ran alternatively within the simulation environment at Dedalus laboratories in Livorno (eight months of almost continuous up-time). Tests have been carried out from early April to December 2011 for nearly 3000 hours of fleet uptime and nearly 5000 hours of AmI Core server uptime. More than 2 million HCL messages and notifications have been exchanged within 24 to 72 hours lasting successful tests. In all the field demonstrations the AmI Core and the on-board robot SWs proved to be crucial in order to achieve the expected results. In addition to several internal testing sessions held in Livorno (both in laboratory and on sea), the AmI Core was demonstrated in the following public demonstrations:
Jun 23th-24th, 2011, Marano (Venice-Italy))
Jan 11th-12th, 2012, Most na Soci (Isonzo-Slovenia)
Jan 28th, 2012, Livorno (Italy)
In conclusion, The AmI platform is:
? modular and adaptive, based on software agents cooperating through a communication bus according to peer-to-peer pattern;
? natively distributed and scalable, by allowing new agents machine to be added to the communication infrastructure;
? open and portable, implemented with open tools (XML) and open source software (RDBMS, Communications), service oriented approach (eService) and portable programming languages (Java).

The HydroNet platform SW is a complex, distributed and heterogeneous AmI platform able to automatically sense when resources, software components or communication networks need to be reallocated or re-configured and to perform the necessary adjustments reducing the human interventions. By considering the project objectives, the AmI Framework assures that all the transactions, data analysis and resources will be not tampered and that the results will be trustworthy.
These features assure innovative advantages of the HydroNet platform respect to traditional environmental monitor systems, such as sensor motion capability, high flexibility, self-organisation and dynamic reconfiguration capability of the system, high scalability, distributed and pervasive water monitoring (global, in terms of space and time, distribution of interesting pollutants and mapping capability), pollutants sources finder (e.g. along rivers, in lake, etc.), survey and preservation of ecosystems and environmental resources.
All the AmI SW has been packaged and released in a distribution CD with installation and usage manuals.

Lessons learnt and conclusions
High detailed monitoring site maps, initially uploaded into the AmI or dynamically provided by an Internet services, are fundamental to program, monitor and review HydroNet missions. When travelling for snapshot demonstrations or campaigns (e.g. in Most Na Soci), if access to the WEB (at least for a significantly sized area) is unavailable, the initially uploaded maps are essential. Not to change too much of the AmI structure of the application, the initially uploaded maps should be organized as local, internal service wholly similar to that currently available on Internet sites. In the today open source world, such a tool does not exist and it would be interesting to check if any tailoring could be possible. For security and accuracy reasons, the need of professional, certified, high precision maps is fundamental in the standard HydroNet scenarios; normally the floating robots should monitor areas very near to the coast or to barrier or very narrow as lagoon or rivers. This maps kind are today not-available on the free market and commercial agreements could be signed with very specialized companies operating in the pleasure nautical market.

In order to obtain excellent visualizations, AmI user interfaces use two different instruments: a common WEB browser with Flash support and Google Earth for 3D visualizations. In order to have a more compact and user friendly experience it would be interesting to merge into a single application all the functionalities (tasking and 2D monitoring plus 3D advanced monitoring and analysis) while still keeping the application a Rich Internet Application. This could possibly be achieved by exploiting the new features of HTML 5 and in particular: Canvas, 3D and Web Sockets.

The field tests and demonstrations showed the need of an user interface monitoring mission oriented. Currently, the HydroNet user interface can be defined as navigation oriented born from the initial need to follow the develop of the entire platform. It may be worthwhile to introduce an user interface more focused on trails and paths. By handling routes and paths as central elements of the mission, event-based commands like sampling or communication or configuration could be nested into the navigation based structure.

In the HydroNet project, Dedalus designed and developed an Ambient Intelligence platform able to support a fleet of floating robots and buoys in order to monitor diverse water bodies for the discovery and tracking of heavy metal and hydrocarbon pollution as required by the initial HydroNet project requirements. The AmI platform is composed of user friendly interfaces for task definition and scheduling, as well as fleet surveillance features which we consider to be effective and innovative in many aspects. The AmI platform stores realtime measurements gathered by robot on-board SDs and it is able to present them in different ways from plain text files to very sophisticated 3D graphical and geo-located views, and to accomplish the requirement of supporting operators in both the short term and long-term analysis of the acquired data.

Important results have been achieved by integrating the UOL's data modeling tools into the Decision Support Service of the AmI platform. This integration has been used for the guidance of a fleet towards a potential pollution source, the generation of alarms depending on realtime and historical data and for the fusion of different data sources into a 3D rendering module for augmented information visualization.
Notable results have been obtained by developing a full-fledged Executive Layer module and two Behavioural Layer modules running, respectively, onboard of the robots and on the buoys. In particular the breaking through choice of adopting the dynamic scripting language LUA made it possible to smooth out several critical aspects related to maintenance, stability and communication protocol definition.

Thanks to the HydroNet experience, the Dedalus team could increase the technological awareness and knowhow in domains such as embedded computing, geo-referenced information handling, tasking of intelligent agents, use of statistical models to generate simulations and decision support services. It has been also very challenging to learn on how to optimize and to reduce the time-to-demo factor for a moving demonstration site.
4.1.3.5. HydroNet PowerPontius radio communication module
The AmI SW control station is the core of the communication infrastructure based on radio communication modules (PowerPontius) with 2.5W output power on 434 MHz with 12V power supply developed by HSLU. The SW for the radio module implements a stand-alone multi-hop routing protocol. The implementation is very close to the Babel protocol (RFC 6126: The Babel Routing Protocol, J. Chroboczek, PPS, University of Paris 7, April 2011, ISSN: 2070-1721), which was developed as a future protocol in MANets. This approach was chosen after evaluating available multi-hop routing protocols and a buy-or-make decision.

Min Typ Max
Supply Voltage 11.75 12 12.75 V
Power consumption receive @12V Vcc 4.5 5 mA
Power consumption transmit @12V Vcc 420 500 mA
Temperature range -25 85 °C
Output power @12V Vcc, 25°C 33 dBm
Output power (-25°C 85°C) 32.7 33.3 dBm
Output power (11.75V 12.75V) 32.7 33.3 dBm
Efficiency 33.3 39 %
Sensitivity @ 9.51 kBit/s -89.5 dBm
Data rate 1.2 38.05 78.1 kBit/s

Table XIII - The HydroNet PowerPontius electrical specifications.
The considerations, adjustments and simplifications made to Babel to run in TinyOS environment are described in the paper Babel Multi-hop Routing for TinyOS Low-power Devices, written by HSLU and presented as reviewed paper at the UBICOMM 2011.
The PowerPontius V1.1 is based on the TinyNode584. For the HydroNet project adapted to HydroNet requirements. PowerPontius is able to transmit data over larger distances and its interface to the PC is RS232. The PowerPontius electrical specifications are are:
4.1.3.6. HydroNet cost- benefit analysis
A cost- benefit analysis of the HydroNet platform was conducted. The main advantage provided by HydroNet Project was from the fact that HydroNet floating robot-boats are autonomous in performing water sampling and analysis, allowing eliminating traditional procedures with a consequent saves of time, resources, and costs. Moreover, since it is autonomous, HydroNet platform allows defining a more precise, distributed, pervasive and timely intensive monitoring plan. The robots can execute a predefined monitoring plan without or with partial human intervention and they can operate continuously by performing analysis in the same place or in different places at defined time, allowing also the detection and prediction of risk situations or to find pollutant sources. This allows performing better water quality analysis and thus a better preservation of the environment. It is difficult to economically assess these further, likely benefits because current, traditional monitoring procedures do not furnish pervasive, on time data. In summary, quantitative and qualitative evidence is pointing towards net benefits from the HydroNet platform that fully justifies its industrial implementation.
As further described in section 4.1.4.2 the contacted potential end-users, mainly state environmental agencies, always confirmed their interest to use the HydroNet platform as a further powerful tool in their monitoring campaigns but they always underlined the need to upgrade and to try, on their own, the HydroNet platform for a limited interval time, 3-4 months. All the HydroNet Consortium efforts to find an industrial or services provider partner, which should be a large company, to produce a small number of industrial prototypes of the HydroNet platform did not succeed until today. The EU national states central administrations and the EU commission should urge the adoption of these new robotized monitoring tools through appropriate programmes and financial instruments; in other words, they should be the engine of a structural change for environment monitoring, water in particular, like eco-energy programmes.

Potential Impact:
4.1.4.1. Project potential technical impact
In summary, the HydroNet platform is a new tool for water monitoring operators to perform water sampling and analysis in a more viable and cost effective way. The HydroNet platform is intended to partially substitute traditional procedures which currently are performed according to the following practices:
1. a boat with some operators on-board leave from the harbour and reach different sampling locations; in each sampling location, the operators use a sampling device (such as Niskin Bottle or Van Dorn Bottle) to take water samples at different depths, transfer the samples to special containers and store the containers on the boat.
2. when back to the harbour, the operators transfer all the container to their labs;
3. in the labs, specialised operators analyse the water using lab equipment devices;
4. results of the analysis are published or stored in digital form in database for consultation.
This procedure involves many operators and requires time and resources such as the boat and other equipment. More details about traditional monitoring campaign are reported in Deliverable D7.5 Market Analysis with Model of the Market.
This procedure involves many operators and requires time and resources such as the boat and other equipments. More details about traditional monitoring campaign are reported in Deliverable D7.5 Market Analysis with Model of the Market. As an example, we report here water monitoring procedure performed in Tuscany by ARPAT (Agenzia regionale per la protezione ambientale della Toscana), the Regional Agency for Environmental Protection of Tuscany.
From 2010, along the 400 km of coastline including islands, 19 sampling points have been identified (see Fig. 45); samplers are taken every two months at 500/1000m from the coast. In the monitoring days, a boat (the Poseidon) starts from the port of Livorno, reaches a sampling point and collects water at different depths (up to 50m). Depending on the distance the boat can move to different sampling points. Samples are bringing back to the port and from there transferred to the laboratory for analysis.

Potentially, the HydroNet platform intends to replace all these phases by providing immediate results of the status of water body for consultation. The basic advantage provided by HydroNet platform is its autonomy in performing water sampling and analyses, allowing eliminating traditional procedures with time, resources and costs savings. Moreover, HydroNet platform allows to define and execute a more precise, distributed, pervasive and timely intensive monitoring plan with respect to traditional procedures, executed autonomously by the fleet of robots and buoys without or with partial human intervention.

HydroNet platform can impacts on the water technology industry specialised in the manufacture of monitoring systems that will increase their market shares thanks to the manufacture of the new integrated system adapted to the market needs. Measurements of pollutant concentrations are also strongly needed in other industries, for example close to installations as petrochemical and chlor-alkali plants, pulp industries and waste handling and treatment; to determine whether legal standards are complied with; and to gain knowledge and understanding that can be used to improve modelling, prediction and mitigation of the effects of pollution.

The application of the platform is not limited to the heavy metals and oil monitoring: the platform is applicable to a large number of environmental applications and pollutants simply by changing the environmental sensor module using a well defined Plug aNd Play interface.
Application of miniaturised sensorised robots to environmental monitoring should decrease costs of monitoring and give benefits associated with new applications. Traditional environmental monitoring systems are static, their application to dynamic monitoring is very expensive.
Water cruises with goal of mapping of spatial pollutant distribution are expensive (expenses for employees and exploitation of ships). The network of buoys and floating sensorised mini-robots are automatic (without human participation, install&forget), with low power consumption and costs of maintenance. Mini-robots give possibilities to achieve tasks impossible to do with traditional systems (e.g. pollutants sources finder, distributed and pervasive environmental monitoring).

The potential impact of the HydroNet achievements spans over the huge sector of water bodies management services, a worldwide problem that involves developed and developing countries. The potential impact is huge mainly in the dramatic modifications of the technologies used to assess the chemical and ecological status of water bodies.
Water is a heritage which must be protected, defended and treated as such. HydroNet platform radically changes the current approach by supplying decision makers with new and more effective, ubiquitous systems and technologies aimed at spatial, temporal, real-time, in-situ water quality assessment. The massive introduction of sensorised robots and ICT technologies in water environments, for improving water quality assessment and for better managing this inestimable resource, would bring conspicuous and relevant advantages, both from a health, social and an economical point of view.

The policy-makers and local authorities in several countries can be direct users of the outcomes of the project. Different pilot sites with different features have been selected in order to demonstrate the feasibility and functionality of the HydroNet platform to citizens, policy-makers, environmental operators, personnel of local municipalities and to any other entity, public or private, that could be interested in the HydroNet project.
The HydroNet project contributes to the impact expected from the EU programme in which it was proposed, which is intended to promote sustainable management of the environment and its resources through developing new technologies, tools and services, in order to address in an integrated way global environmental issues. HydroNet proposes in fact an integrated approach to water monitoring. The sensorised robots network proposed in HydroNet has a wide applicability, to provide a unified response to the questions related to water monitoring, not at a local level, but a global level. On the other hand, the problems of water monitoring, management and preservation cannot be approached at a local level, intrinsically. Waters span over the entire globe and affect the life of the whole population.

In the selected Italy and Slovenia validation sites, the HydroNet platform has been set up to demonstrate how rivers, lakes and coastal waters areas can be easily monitored with a sensorised robots network furnishing a wide range of physical and chemical parameters as stated in the Water Framework Directive. These demonstrations have been performed in collaboration with the local environmental Organisations and Municipalities, which represent important decision makers and potential users of the HydroNet platform and methodology.

Results obtained by the pilot site activities have been crucial from a technological perspective and also from a social viewpoint because they allowed evaluating the sustainability of the proposed solutions, as well as the acceptability by decision makers towards a new integrated approach to the problem. The results achieved in the project can have a direct impact on very important areas:
- water preservation and monitoring
? the environmental SDs embedded in the autonomous and cooperating robots allow a monitoring in real-time, in-situ of a wide range of parameters (physical and chemical), supplying information on the chemical and ecological status of water bodies (e.g. rivers, natural and artificial lakes, coastal waters, etc.) and providing the bases for management tools for sustainable water systems;
? the use of robots allows to carry out cost-effective monitoring campaigns every time it is necessary (also during the night), and not only in relation with the workers shift as it happens using traditional technologies. The increase in time and in space of monitoring activities can assure a consequent improvement of surveillance of the healthiness status of the water resource.
- safeguard of the human health
? the sensorised networked robots, monitoring spatially and temporally the water quality, can represent a scientific method providing punctual information on the presence and distribution of highly toxic elements, such as cadmium, chrome, mercury, which, for their proprieties of bioaccumulation and biomagnification along the food-chain, may be considered a serious threat to human health;
? the HydroNet approach is aligned with the interesting Health Society initiative promoted in Tuscany (Italy) by the Regional Healthcare Program (RHP). Nowadays, the Health concept involves different aspects of the human life, and the absence of diseases and illness is only one (obviously the most relevant) of them; other important factors, such as the lifestyle and the psychological, social and environmental wellbeing, represent important indicators to evaluate the global health status of citizens. For this purpose, the RHP proposes, through the Health Society initiative, the integration between the policies carried out by the Healthcare, the Social and the Environmental Systems. In this new interesting scenario, an important role is played by the environmental policies, especially in the context of urban areas (e.g. the number of motor vehicles, the noise threshold, the presence of parklands and open spaces, etc), the energy consumption, the quality of the air and the air pollution level, the management of the waste and, obviously, the quality of the water bodies. The HydroNet integrated platform is really a very promising solution to address these issues. In particular, the modularity and the scalability of the HydroNet platform, based on the Ambient Intelligence paradigm, will be able to offer specialised services to different actors in charge of the water, social, economical, environmental and healthcare management. Moreover, some partners of the HydroNet project are studying and experimenting different approaches for the integration between an ambient intelligence system and the clinical software modules of the Healthcare Information System.
- Risk prevention
HydroNet platform provides an integrated solution to water monitoring. Monitoring is the primary tool for risk prevention. The HydroNet platform is be able to monitor a wide range of waters in a spatially capillary way and in a temporally continuous way. This accurate monitoring allows early detection of pollution risks, as well as risks for the deterioration of water quality, due to whatever reason.
- Social services
As stated, water is a fundamental resource for humanity. Monitoring and preserving water is a service provided to all. By supporting sustainable management of water, the HydroNet project provides social services indirectly to all citizens. The local administrators who will take advantage of the HydroNet network on their local area will offer to their citizens the primary service for their health, well-being, food and wealth.
- Economic Impact for the Microsystems, Information and Communication Technology (ICT), and Robotics Industry
? the HydroNet project exploits the recent advancement of microsystems, robotics and information and communication technologies, proposing an integrated platform for the water quality assessment to enter the market of technologies and equipments for the decision makers and for any other Institution involved in the water bodies management.

The new distributed intelligence platform proposed by the HydroNet project is a full answer to each of these aspects. That is why the strategic impact can be very large.
The HydroNet platform is designed for a massive application, which is for being developed in large numbers at relatively low costs. This can have a positive economic impact on the European ICT industry. At the same time, HydroNet can provide tools for an advantageous cost/benefit ratio in public services of water monitoring, management and preservation.

HydroNet platform works as multi-sensory robot networks with communication capabilities able to monitor water bodies, such as rivers, natural and artificial lakes, coastal waters, etc., not reachable with standard monitoring systems. Robots can also penetrate areas not accessible by big boats or land-based systems. This is particularly true for changing environments like rivers, which can modify on seasonal basis their water level, alternating flood with dry periods, or lacustrine areas difficult to reach. Moreover, HydroNet robots is able to find a pollutant source and to map an area with spatial and temporal pollutant distribution.

The multiscale transport and transformation of heavy metals and oil, detected by the HydroNet systems, can play a significant role in the definition of the water quality and the improvement of knowledge on this phenomenon carries economic and social impacts. This improvement is achieved through integrated solutions based on robots, miniaturised sensing systems and wireless network technology for the water quality assessment and risks management posed by heavy metal pollutants and oil to ecosystem and human health, as well as to relevant economies (fishery, aquaculture and the quality of life and environment).
Heavy metals and oil are considered pollutants with high risk of toxicity, persistence and bioaccumulation. Therefore, knowledge of concentration, transport, and dynamics of heavy metals in water bodies is needed to predict the potential impact on human as well as aquatic life. Thus, contamination is not only just a local issue but also has global dimensions.

The decision makers and local authorities in several countries can be direct users of the outcomes of the project and can plan future economic development and policy-making toward meaningful and cost-effective monitoring robotised systems.
4.1.4.1.1. HydroNet platform
The opportunities to exploit the HydroNet platform are closely related to two factors: first, it should be demonstrated, especially to final users and customers, that HydroNet provides an effective solution for water monitoring with better performance, or at least comparable, to current monitoring procedures and at reasonable costs. Second a strong push to the use of this innovative solution should be provided by decisions maker and/or legislators which could force the adoption of innovative and high-tech solution that improves water monitoring with respect to currently available practices.
Above all, a necessary condition for future and commercial exploitation of the HydroNet platform is a testing and validation of the system in a real scenario with the collaboration of end-users. The testing should aim at evaluate both technical and functional performance of the platform as it is now in a real settings and in real conditions. The validation should:
- assess the current technical performance of the robotic platforms and of the SDs;
- assess the performance of the integrated platform in terms of effectiveness, reliability, robustness and usability;
- asses the benefit provided by the platform for water monitoring and how the system meets the requirements of end-user and WFD
- provide indications on how tuning the system on user needs;
- provide fruitful indications on how to improve the platform from a technical point of view and in terms of the service provided;
- provide indications on other possible usages and applications.
Moreover, this setting should be considered a permanent test bed for continuous improvement and testing and to demonstrate to potential users and investors the potentiality of HydroNet.

In synthesis, the plan for the exploitation of HydroNet requires:
- to improve the single components of the HydroNet platform;
- to complete the development of the SDs, including their validation;
- to search for financial support for the testing and the validation of the HydroNet platform at national and international level by submitting research proposal to the European Commission or to local government support research programmes.
- to sign an agreement with an agency for environmental protection and coast guard to set up the HydroNet platform in a real monitoring area;
- to perform a real water monitoring campaign under the supervision of the agency for environmental protection.
- to perform more extensive dissemination activities to promote the HydroNet platform using the permanent test bed as demonstrator.
The improvement of the current SDs in terms of sensitivity and the development and integration of new electrochemical, optical and physical SDs for measuring other pollutants as defined in the WFD should be considered.
4.1.4.1.2. HydroNet Robots
The HydroNet robots are at the state of the art of robotic technology and their capabilities show the possible use of robotic autonomous agents in different scenarios, especially in environments inaccessible to human operators or in situations where the robots can provide effective methods to explore/monitor large areas without or with a limited human intervention. For these reasons, robotic agents will be used more intensively in the future in water environments.
Underwater robots (AUVs Autonomous Underwater Vehicles) have been already widely developed and studied, and the continuous improvements in SDs, power supply and actuator technologies have given the possibility to use them for more and more challenging tasks. As few examples, AUVs have been developed for operations in shallow water, for localizing chemical plume and for finding mines. More recently, the use of a AUV fleet has been proposed to guarantee a larger area of operation, to fully map some parameters and to assess spatially and temporally water quality. Several underwater gliders are used in California and research works allowed robots able to communicate and to cooperate in water environments. Of course, AUVs can be used together with SDs networks in fixed position. The SDs can measure some parameters and call the mobile autonomous robots in the areas of particular interest.

HydroNet project confirms that networks of SDs and robots can move also on the water surface to guarantee an efficient and effective monitoring of rivers, lakes or coastal environments quality.
The analysis presented in Deliverable D7.5 Market Analysis with Model of the Market and in previous sections of this document well evaluated the strengths, weaknesses, opportunities and threats of the HydroNet platform and especially of its autonomous surface vehicles (ASV) for water quality monitoring according to WFD. To summarize, HydroNet ASV fleet is an innovative instrument based on ICT and robotic technologies, but still many actions have to be implemented to adopt it as a real alternative to the environmental agencies current approach.

General analyses have been carried out by taking into consideration the goal of the HydroNet Project which allowed the ASVs realization, but these new ASVs can be used in many other activities (by adding new tools), such as:
- to map the seas, rivers, lagoons beds with video cameras, multi-beam sonars,
- to map galleries layouts,
- to take sediments samples of water sites,
- to clean water bodies from small objects such as plastic bottles,
- to clean water bodies by applying bioremediation or solvents,
- to support unmanned underwater vehicles in the underwater navigation by sending acoustically localization data,
- to work as a communication relay for underwater vehicles,
- for port protection by surveying an area with video cameras.

Exploitation of the HydroNet ASVs fleet needs the following steps:
- raising funds for additional activities to strengthen and optimize this research prototypes,
- finding a private/public industrial partner to be involved in the production of an industrial prototypes,
- industrial prototypes commercialization and selling.
HydroNet Consortium Partners are trying to raise more funds for the first step from national and European projects. The second step will be more difficult and requires analysis of the most promising applications interesting industrial partners. In a near future the market will ask for such sustainable product: this is a todays opportunity and also a concrete difficulty in this recession period.
4.1.4.1.3. HydroNet SDs
4.1.4.1.3.1. Electrooptical SDs
Lumex considers AE-2 Mini and CRAB SDs as commercial products. The latter, already fully operational, can be used both on-board and at stationary points. Further improvements will be made to adopt both SDs to platforms which have hydraulic and power supply subsystems different from HydroNet platform.

The AE-2 Mini SD has been submitted to State Harbour inspection (Moscow and St. Petersburg) to equip two vessels for coastal water (Gulf of Finland) and rivers monitoring. A modified Fluorat AE-2 Mini online water monitor for water treatments plants and industrial applications is under development. The core element of this new TPH water monitor is the miniaturized optical and extraction module developed for the HydroNet AE-2 Mini SD. The new model will be introduced to the market in the 4th quarter 2012. The HydroNet AE-2 Mini SD is now being modified to a portable handheld version (named PIRN) and will be used for sanitary and environmental monitoring waters. Commercial release is planned in the 1st quarter 2013.

Numerous proposals had submitted concerning the CRAB SD and Lumex has got an order for manufacturing of several units for State Harbour inspection with coastal vessels (Moscow), for river water monitoring (St. Petersburg, Moscow, Ljubljana), for waste water control at refineries (Lithuania, Germany, the Netherlands, Thailand, Malaysia and other countries). RT and SSSA have plan to propose CRABs for coastal water and Arno river control in Livorno and Pisa to the corresponding local state administrations.
4.1.4.1.3.2. Electrochemical SDs
An application for further dissemination of the electrochemical SDs developed by HUJI has already been submitted to two agencies in Israel: the Applied Funds of the Hebrew University and the National Agency for Sewage and Wastewater. Recently, the application to the Hebrew University was approved. It is expected that an additional research for 1-2 years will end with a commercial prototype. The utilization of this generic platform, which was developed by HUJI, is very attractive for remote monitoring of drinking water and even more waste waters. There is currently no way to constantly on-line monitor wastewaters for heavy metals. Many industries, e.g. paints, military and plating, use hazardous metals and their wastewater is only sampled occasionally. It is anticipated that the SDs will allow monitoring specific metals and warn once their concentration in the effluent exceeds a predefined value.

The cadmium SD has several significant advantages:
1. The sensitivity of the SD is 50 ppt, which grades it as one of the most sensitive SDs for Cd, competing well with very sophisticated techniques such as ICP-MS.
2. The dimensions of the SD are very small and so is its weight. This allows high portability and implementation in different environments including carrying by autonomous vehicles such as boats.
3. Totally automatic including calibration and transmission of data to a central control system. This has an enormous advantage for the continuous and autonomous monitoring of remote and hazardous places such as nuclear plants.
In view of these advantages we are currently trying to further implement this SD for the remote sensing of water sources such as well and sewage waters. It is evident that this and the other SDs can be used in different applications where real time monitoring of heavy metals is crucial, such as in controlling the quality of drinking water.
The electrochemical cell, which we designed for the sensor is new and will be patented soon.

The Hg sensor is of particular interest due to the high toxicity of this metal and its natural abundance and use by man. The heart of the SD is either gold or glassy carbon electrodes that are part of the same flow system used for the Cd and Cr SDs.
Exploitation of the Hg SD will have to go through two steps: first raising funds for an additional year for validation and final design of the system. The second step will involve commercialization and will require a SME or another industrial partner to be involved in the production of an industrial prototype. HUJI will be able to raise the funds for the first step from a national base agency. The second step will be more difficult and require analysis of the best and most promising applications, e.g. wastewater, drinking water or environment. The final prototype will have to be designed for a specific application.

The SD for chromate is similar to that of the cadmium in respect to the flow system, yet, the chemistry is quite different. The advantages of these SDs are:
1. The ability to sense only Cr(VI) which is the most hazardous species of chromium. The standard methods are not sensitive to Cr(VI) only but measure the total concentration of Cr.
2. High portability due to its small dimensions and weight and high sensitivity that allows it to determine quite low levels of this metal.
3. Low cost and autonomous operation.
As for the Cd sensor, HUJI aims at applying this sensor for the remote sensing of drinking water and more importantly for sewage water. The latter is of outmost importance in various industries such as paints and dyes, military industry, etc. HUJI is seeking funds to continue this work. The flexibility of these systems that are based on the same flow system and electrochemical cell makes it possible to switch from one sensor to the other. Another option is to combine more than one sensor in a flow system and obtain a multi-SD.
4.1.4.1.3.3. Biological SD
The main strength of our product is that is the first automatic device for measuring the bioavailable fraction of mercury in aquatic environments. Such a device does not exist on the market yet. It may not only be used as an integral part of the robots and buoys of the HydroNet project, but can be integrated in different platforms or upgraded to a standalone detector. The weaknesses are related to the biological nature of the measurements, meaning that all the drawbacks of bioSDs, in particular whole-cell bioSDs apply. It needs a skilled operator to prepare the cells and media, which can be automated if commercial exploitation proves successful.
The principal impact of measuring bioavailable Hg is enabling scientists to detect mercury on site in its form that enters organisms. This was not possible before. Additionally, the sensitivity of the bioSD allows measurements in non-polluted environments, which is crucial for understanding the background levels, and more important, to be able to follow spatial and temporal trends. Specific technologies in molecular biology that have been developed, will be exploited separately of the general HydroNet aims and are described in the Exploitation part.
4.1.4.1.4. HydroNet environmental modeling tools
The modeling tools developed in the framework of the project are a valuable tool for determination and prediction of the state of pollution in contaminated coastal areas. Some of the models and modules have already become a substantial part of other national projects undergoing in Slovenia. The established national and particularly international collaborations (e.g. ARPA-FVG, modeling of pollutant dispersion in the Marano and Grado Lagoons and the Gulf of Trieste) will enable further development and improvements of the models and their individual modules.
4.1.4.1.5. HydroNet Ambient Intelligence infrastructure
In order to achieve the impacts for the water preservation and monitoring, the safeguard of the human health, the risk prevention and value added social services, the AmI infrastructure play a basic role being in charge of:
managing the devices used for the monitoring (robots, SDs networks),
acquiring, storing, elaborating and presenting all the information of the HydroNet platform,
interoperating with knowledge management systems,
interacting with all kinds of operators.

Concerning the monitoring of water bodies and the whole environment, the HydroNet project achieved significant results in the specialization of advanced user interfaces and in the configuration, tasking and control of robotic actuators in real environment. It is worth noting that the AmI Core strength can be used with the same effectiveness with human driven boats, currently non integrating sampling devices and, in general, with technologically less evolved tools. The AmI framework could be effectively used in more critical missions, where the data could be acquired with different systems and/or devices.
The AmI SW can be used as distributed and highly interoperable SW infrastructure for the control, monitoring and evaluation of sensorised/robotised environments. It should be underlined that the Dedalus (formerly Synapsis) activities in the HydroNet project belong of a precise company roadmap started in 2004 with the objective to enter the eHealth market by 2010. With the HydroNet project Dedalus completed the experimental activities of pushing the AmI Core SW towards a more pervasive, structured control and actuation framework for robotic controlled environments. The objective is to improve the impacts in the strategic context of remote assistance where it will leverage two of the most crucial aspects of modern healthcare namely continuity of care and integration of different care environments:
secondary care (hospitals, specialist visits ),
primary care (General practitioners, ambulatory accesses ),
homecare.
The AmI Core could become not only a continuous monitoring system for patients with chronic diseases in their home environments but also a system to support integrated clinical and medical workflows in the practice of robotic rehabilitation therapies. The ability of the AmI Core to plan and schedule missions will significantly support the configuration, control and monitoring, even remotely, of diagnostic, therapeutic and rehabilitation pathways.

The AmI Framework includes two important, very interesting components: the Lexicon tool and Algo Service system. Through the Lexicon, it is possible to plan actions, commands, timing and synchronizations in a close-to-natural language while keeping a semantic structure of contents. Dedalus believes that this product has a significant added value in several different systems and services. In general, this service will have a strategic value wherever a semantic interoperability between human operators and computer systems will be required.
Concerning the Algo Service, it should be noted that the use of expert systems able to provide efficient and effective decisional support will become more and more strategic in all mission critical contexts. They will represent a key-feature especially in systems that implement critical workflows; for example in the healthcare sector, the clinical workflow aiming at optimizing both the healthcare practices according to the Evidence Based Medicine and the processes of clinical governance.

HSLU try to find a partner to sell the radio module. Some of the mechanisms of the Babel routing are not yet implemented in the PowerPontius software. Further development will be made at HSLU in terms of student projects. There are also ideas to use WSN test beds (e.g. WISEBED, http://wisebed.eu/site/) to perform more tests.

4.1.4.2. Project potential economical impact

Country/Region Coastline (Km)
Norway 25.148
Greece 13.680
United Kingdom 12.430
Italy 7.600
Denmark 7.310
Turkey 7.200
Croatia 5.835
Iceland 4.990
Spain 7.960
Estonia 3.790
France 3.427
Sweden 3.220
Ukraine 2.780
Germany 2.390
Portugal 1.790
Ireland 1.450
Finland 1.250
Cyprus 648
Poland 491
Albania 362
Bulgaria 354
Romania 225
Montenegro 199
Malta 197
Lithuania 99
Slovenia 47
Bosnia and Herzegovina 20
Gibraltar 12
Monaco 4
Total 114.908
The effect of using HydroNet results in a better monitoring and preservation of the environment. The HydroNet platform can be considered a forward step in for in-situ monitoring, beyond current procedure and legislation, with real-time data telemetry and the capability to produce high resolution distributions in time and space. The HydroNet platform with its real time feature, its integrated dispersion models and its source finding algorithm can allow immediate actions to prevent pollution to extend to wide water area.

Numerous potential uses, ranging from monitoring and measurements in waters surrounding ports, monitor concentration levels from offshore activities, monitoring of water quality in fish farming areas (which is an expending industry), environmental protection through the WFD, as well as in international environmental based conventions and programs, opens for a large potential market. The largest group of costumers and users are first and foremost expected to be national funded agencies in countries with a large coastline and inland waters surface.

Table XIV - The European coastlines.
Weaknesses presented above have not recommended to bring the platform to end users for an independent assessment and, consequently, all the previous benefits have not been verified in the field. In numerous meetings with end-users, in Italy with state monitoring agencies (Lamma, ISPRA, ARPAT, ARPFVG, port police and port authorities) and in Slovenia, after describing the features and functionalities of the HydroNet platform and also performed guided demonstrations, there have been discussions on the use of the hypothetical current platform. Users have always maintained their interest by underlying, however, the following main needs:
a. equipping boats of further SDs as required by WFD and currents campaigns,
b. introducing mechanisms to bring water samples and, if possible, sediments, to the ground station,
c. automating missions as much its possible by also retaining in a database their specifications so that they can simply repeat even after a long period of time,
d. extending the current radio transmission power to avoid buoys use,
e. improving the user interface.

Component Cost
Robot 33.712,00
Winch 1.182,00
Sampling System 6.650,00
Fluidic System 3.000,00
Physical Probe 12.000,00
Total 56.544,00
The cost for the realization of a catamaran robot prototype reported in Table XV, is estimated in about 57K without new HydroNet SDs. Considering that this is the cost of the material for building a prototype, HydroNet Consortium can estimate the production cost of a industrial catamaran robot in about 30K, which include the costs for the materials, the components and the assembling costs. Final price of this robot type could be around 75K (without SDs).
Table XV - Estimation of the HydroNet Catamaran cost.
The realization cost of a flat-boat robot prototype is estimated in about 37K (without SDs). Considering that this is the cost of the material for building a prototype, HydroNet Consortium can estimate the production cost of an industrial flat-boat robot in about 20K, which include the costs for the materials, the components and the costs for robot assembling. Final price of this robot type could be around 50K (without SDs). The way to obtain such as product, that is consolidation activities and the production line set-up, is estimated in about 1.5M with development personal costs.

The cost of the original, custom-made SDs ranges from 5.000 to 8000 (no R&D).
As reported in Deliverable D7.5 Market Analysis with Model of the Market (see chapter 6. Cost/Benefit Analysis), the HydroNet platform permits to reduce costs of monitoring campaigns. It is worth to mention the fact that robots or automatic procedures for water quality monitoring according to WFD are not currently on the market, thus this is a great opportunity to create a new market and labour opportunities.

The one-day mission cost (reported in D7.5) obtained from informal Italian and Slovenian sources and practically, confirmed at end-users meetings, plotted against different market scenarios, shows the economical consistence of the HydroNet platform. Some human resource are still needed to manage the floating robots, others can be redirected to activities with an higher added value.
The contacted state environmental agencies always confirmed their interest to use the HydroNet platform as a further powerful tool in their monitoring campaigns, and they have always emphasized
1. the need to improve the HydroNet platform in agreement with needs above described in this document section and
2. to try, on their own, the HydroNet platform for far enough time periods to find its most effective integration into current environmental monitoring processes and have a genuine knowledge of its overall cost (logistics, operational and maintenance costs).
Although very few potential industrial partners have been met to numerous meetings or workshops and the previous potential large profits have been always presented, no company really interested in building a limited number of floating boats has been identified until today.

HydroNet Consortium believes that the effort required to produce a small number of industrial prototypes of the HydroNet platform is much greater than the possibility of any SME: the financial risk is really critical for SME survival. To start an industrialization process, a major industry operating in the water monitoring should be involved (very few exists and not in Italy) or European governments should make arrangements and funds to steady the new robotic technologies outlined in the HydroNet project. Any case, after obtaining the industrial prototype, to really promote this new market and change the current monitoring approach, the central EU national state administrations and the EU Commission should urge the adoption of these new robotized monitoring tools through appropriate programmes and financial instruments; in other words, they should provide the fuel for a structural change of the environment monitoring, water in particular, like eco-energy programmes.

The whole HydroNet Consortium and several individual partners, as stated in the previous chapters, are trying to raise more funds for the realization of an industrial prototype, its commercial development and the continuation of the research concerning the whole HydroNet platform or its individual components. Except Lumex SME that began the process of industrialization and commercialization of new SDs developed in the HydroNet project, the funds lack will be a serious obstacle to the continuation of the research work.

Also RT confirms its willingness to participate to national and European calls to raise more founds. At the same time, RT is pursuing collaborations with governmental agencies for environmental protection and other end-users to carry out a functional/technical validation of its flat-boat robot and the HydroNet platform. RoboTech intends to make agreements with an agency for environmental protection or similar end-users to identify a site where to install permanently and for an extended period of time the HydroNet robots and buoys to perform real water monitoring campaigns under the supervision and in collaboration with the agency. This setting should be also considered a permanent test bed for continuous improvement and testing and to demonstrate to potential end-users and investors the HydroNet potentialities. Funds for this activity should be held down and probably sustainable by RoboTech and environmental protection agency.

SSSA is a research institute and by using the research prototypes developed in the HydroNet project will try to pursue its mission by proposing projects in the next national and European calls. Deliverable D7.5 Market Analysis with Model of the Market with its cost-benefit analysis clearly shows the high economical potential of the HydroNet approach to the autonomous, automatic, in situ and real-time monitoring in open water: private companies can make a profit, public money can be saved, the whole community can live in an environment increasingly secure, spatially and temporally. Of course, more research activities are still needed to comply with all the EU WFD requirements: they are well above the goals of the HydroNet project but a focused research and development can reach most of them in the near future.
List of Websites:
In the HydroNet WEB site http://www.hydronet-project.eu/ enormous information are available

Project Info: http://www.hydronet-project.eu/index.php?menu=projectinfo
Objectives: http://www.hydronet-project.eu/index.php?menu=objectives
Partnerships: http://www.hydronet-project.eu/index.php?menu=partnership
Pictures: http://www.hydronet-project.eu/index.php?menu=pictures
Videos: http://www.hydronet-project.eu/index.php?menu=video
Newsletter: http://www.hydronet-project.eu/index.php?menu=3newsLetter
in Russia, Norwegian, Slovenian, Italian, German and English languages.
Useful information and legal contact persons and other contact persons are shown in tables contained in the attached PDF file.

Documenti correlati

final1-d1-4-hydronet-final-report-12-09-2012-v2-0.pdf

Final Report Summary - HYDRONET (Floating Sensorised Networked Robots for Water Monitoring)

Documenti correlati

Condividi questa pagina

Scarica