Skip to main content

Enabling Mixed Societies of Communicating Plants and Artefacts


CTI designed and developed an ontology that formalises and represents the knowledge of the PLANTS system. This knowledge is divided into the following three categories: - Conceptualisation of the bioGadgetWorlds concepts, - Characterisation of plants, - Characterisation of sensor/actuator systems, - Definition of rules. The Ontology is organised hierarchically into the PLANTS Core Ontology (encodes general knowledge) and the PLANTS Higher Ontology (encodes application-specific knowledge). Evolution of the PLANTS Ontology is accomplished according to the proposed PLANTS Ontology lifecycle. The PLANTS ontology provides a conceptualisation of the bioGadgetWorlds, enabling the use of this knowledge from other systems. This conceptualisation ensures the semantic interoperability among eEntities and supports a service-discovery mechanism. The knowledge represented by the PLANTS ontology supports the decision-making process that is necessary for the PLANTS system. We have also specified how the ontology could be enabled to learn new knowledge by giving the detailed specifications of a machine learning application, which uses the plant’s processes (e.g. photosynthetic process) to determine optimal growing conditions. The PLANTS ontology is written in DAML+OIL and developed with the Protégé-2000.
Having installed ePlantOS what is necessary is to define the way that the hardware (biosensors/bioactuators) are going to work with ePlantOS in order to provide data in the case of sensors, or accept commands in the case of actuators. To achieve this, the definition of an ePlantOS specific driver for each sensor or actuator is required. The driver specification depends on the interface between software and hardware; In the case of sensors grouped by an FPGA board or PIC microcontroller the manufacturer can define the hardware’s output himself, while in the case of a Commercial Off The Shelf (COTS) sensor device the output is predefined by the company that builds the device. In order to build the driver the first thing to do is implement a Java interface, which determines the information exchanged between the ePlantOS and the hardware. For one to build a specific driver protocol, it suffices to produce a new implementation of this interface. Inside this implementation the developer of the driver has to write the suitable commands to communicate with the sensor/actuator and forward the information acquired to the ePlantOS.
One of the most important features of the PLANTS system is its decision-making process that is composed of the following three levels: - Diagnosis of the plant state, - Local decision-making, which determines the possible actions of an eEntity, - Global decision-making, which is used when a global state must be preserved/detected. All these levels of the decision-making process are based on a set of rules that determine the way that a decision is taken and must be applied on existing knowledge represented by the PLANTS ontology. In addition, CTI has explored the application of machine learning in PLANTS system for inducing new plants domain models (typically in the form of rules or decision trees) in order to extend PLANTS Ontology with new knowledge. New knowledge is produced by monitoring and analyzing a combination of sensor data with two or more sensors.
The purpose of this methodology is to describe the process that one should follow in order to transform a regular plant into an ePlant. This process may be called plant virtualization, as it provides the steps to create a virtual (computing) space to be integrated with the physical-biological element. We have described these steps of the plant virtualisation process ranging from enhancing the plant with computational and sensory communicational abilities, to installing the required software that will enable it to participate in various applications. A driver-based mechanism has been specified and a methodology to develop efficiently I/O drivers has been proposed to support effective interfacing with sensor devices. Furthermore a driver configuration approach has been specified to facilitate the deployment and integration of the I/O drivers.
A hardware module that is based upon an FPGA processor with digital and analog I/Os allows for communication to a variety of different commercial sensors. The system allows for analog devices that have voltage, current or resistive changes and can also relay or modify digital signals for control of electronic based sensor systems.
This relates to the actuator systems that have been put in place for the PLANTS project. They are commercial irrigation systems that have been modified by the addition of controlling units that make them part of a distributed system. This allows great flexibility and control to be given to the system.
The concepts of the PLANTS project have been demonstrated to the public through a workshop, open for 4 days during October 2004 at the Eden Project, Cornwall, UK. Evaluation of the effectiveness of this dissemination activity in communicating the aims and principles of PLANTS was carried out by a number of methods. Attraction power, a measure of the number of visitors to the workshop against the total number of visitors to the Eden Project, was calculated. Within the workshop the route taken through the exhibits by a random sample of visitors was mapped. This visitor tracking data shows which elements of the workshop/PLANTS project attracted the most attention and how visitors interacted with the information displayed. A preliminary data set was drawn up from personal meaning mapping, a technique that charts understanding of the key workshop concepts pre and post visit.
A software module, based on GAS-OS, that enables the formation and management of synapses between eGadgets and ePlants. ePLANT-OS provides an interface layer with ePlant sensors and actuators, maintains an enhanced, plant-specific ontology and supports distributed resource management with local and/or global decision-making according to criteria. ePlantOS is a middleware system operating according to a peer-to-peer interaction model in which plants being monitored and devices of the system are universally modelled. A modular system design allows the replacement of a module without affecting the functionality of the rest. The modular design of ePlantOS allows the integration of up-to-date algorithms and protocols in the form of plug-in modules. Finally, the system incorporates machine learning mechanisms that allow the extension of the PLANTS ontology with new knowledge in the form of new rules.
This is a method to determine whether a plant would be best suited to shade or light conditions. Generally speaking domestic plants are bought from garden centres and nurseries and the buyer is dependent on the producer informing them as to what conditions would suit the plant. Very often the plant is not grown in these optimum conditions due to space and production pressure constraints. Hence shade and light plants bought commercially can give Electron Transport Rate (ETR) readings that are not distinguishable as they have adapted to less than favourable conditions. For this method, when plants are bought they need to be conditioned to either shade or light conditions. Using chlorophyll fluorescence an ETR curve is acquired. This is then analysed to determine the inflection point of the curve (inflection Photosynthetically Active Radiation (PAR)). If the inflection PAR is greater than 200 PAR then the plant optimum conditions would be in a sunlight area. If the inflection PAR is less than 200 then the plant is best suited to a shaded area. Interior and exterior landscapers could use this method to determine the best positioning of plants.
In the PLANTS project CTI designed and developed a set of graphical tools that provide the user with a set of useful operations like creating a bioGadgetWorld (mixed society), viewing knowledge represented into the PLANTS ontology, monitoring the ePlants that take part in his/her bioGadgetWorlds and finally managing the rules taking part in the decision-making process. The first tool is the bioGadgetWorld Editor (bioGWEditor) that can be used even by non-expert users for the creation and editing of bioGWs. Two graphical interface versions of the editor have been created, to test the primary editing functions: on a laptop PC and on an iPAQ. Editor functionality: - Discovery of surrounding devices and plants and the capabilities they offer for interconnectivity. - Information (visualization) of the above discovered item. - Supporting Users in creation, editing and destroying association links (Synapses). - Creation of Functional sets of links, that serves user purposes (a defined collection of associations). - Operation (activation, deactivation, elimination) and management (i.e. editing) of each functional set. - Functionalities similar to the ones of the SLADA tool targeting the monitoring of a bioGW and the management of the global decision making process. The second tool is the Ontology-based Supervisory Logic And Data Acquisition (SLADA) tool that provides the user with a set of useful operations for monitoring and manipulating the information for a specific eEntity. Specifically the SLADA tool is a graphical user interface, which provides the user variable operations for viewing the knowledge represented by an eEntity’s ontology, monitoring an eEntity and managing the rules of an eEntity local/global decision-making process. Finally the Rule editor is a tool that provides a Graphical Design Interface for managing (Create/Retrieve/Update/Delete) rules, based on a user friendly node connection model.