Final Report Summary - LOTUS (Localisation of threat substances in urban society)
Executive Summary:
Detection of explosives has traditionally been focused on detection when an explosive charge is already ready for use, being transported to or already at the scene of an attack. Preventing terrorist attacks while they are already in motion is extremely difficult. Not only is it difficult to find an explosive on a person or in a bag but one must also face the need to do this in a very short time and then immediately implement measures for countering the attack once a possible suspect or suspicious object has been found. The capability to intervene at an earlier stage is therefore of very high importance. The preparation and production phase is by necessity quite long. Unless the bomb-makers are quite professional they will need to test and practice the procedures and in many cases the chemical manufacturing process in itself takes considerable time. The LOTUS system can be used during primarily the production phase in certain scenarios.
The LOTUS concept is to detect precursors over a wide urban area. The detectors may be placed at fixed positions although most detectors should be mobile. The size of today's detectors makes placement suitable in vehicles such as a police or other law enforcement vehicles. These distributed detectors continuously sample air while its carrier performs its daily work. The driver of the carrying vehicle does not need to operate the detector and in fact, should not even know if any threats are detected. When a suspicious substance is detected in elevated amounts, information about the type, location, amount and time is registered and sent to a data collection and evaluations centre for analysis. Several indications in the same area will trigger an alert, enabling law enforcement agencies to further investigate and respond. The LOTUS system should be regarded as a technical tool for intelligence gathering, rather than a detection system, meaning that the information obtained should be combined and confirmed with information from other sources to make the most accurate assessment of the situation. LOTUS is a project where a system concept has been developed and tested.
The outcome of the research work has produced softwares, hardwares and knowledge on the requirements for the detection of production of home-made explosives and drugs. The research work in the LOTUS project has encompassed e.g. softwares for sending and encrypting data containing position, time and sensor data over the mobile phone network, development of computer algorithms for evaluating sensor data with respect to reducing false alarms and backtracking sensor data to potential release sources (bomb factories) and dispersion modeling calculations in order to understand the dispersion properties of threat substances in vapour phase as released from an indoor laboratory to the nearby outdoor environment. Furthermore, the development of detection techniques and the integration of these into sensors that could make it possible to use them on mobile vehicles in urban areas have made it possible to test the entire LOTUS system concept in real field trials in three capitals in Europe; Stockholm, Finland and Madrid.
It can be concluded that the proposed LOTUS system concept, using mobile automatic sensors, data transfer /location via GSM/GPS for on-line detection of illicit production of explosive or precursors to explosives and drugs is a viable approach and is in accordance with historical and todaya's illicit bomb manufacturing.
If a LOTUS system had been existing 2005 and had been put into operation before the London bombings, the Leeds bomb factory could have been discovered. Consequently, this would have offered police forces the possibility to intervene and catch the terrorists at an early stage of the terrorist activities.
Project Context and Objectives:
One of the largest scientific, technological and deployment challenges in the security domain today is the development of systems for localisation of home-made explosives manufacturing facilities or prevention of the use or enhanced detection possibilities of commercially available precursor chemicals. The improvised explosive devices, that can be prepared in ordinary kitchens, contain explosives that in terrorist attacks give devastating effects. For the law enforcement services there is at present an urgent need for tools that contributes to the early discovery and neutralisation of illicit production of home-made explosives.
The most important and concrete contributions from systems such as LOTUS with respect to the complex task of locating bomb factories or hindering the production are:
- An early warning
- Localisation of the bomb factory
- Automatic system
- Less personnel requirements as compared to normal intelligence
The concept of the surveillance systems for the localisation of IED illicit bomb factories is unique in this area as there today are no commercial systems available.
Detection of explosives has traditionally been focused on detection when an explosive charge is already ready for use, being transported to or already at the scene of an attack. Preventing terrorist attacks while they are already in motion is extremely difficult. Not only is it difficult to find an explosive on a person or in a bag but one must also face the need to do this in a very short time and then immediately implement measures for countering the attack once a possible suspect or suspicious object has been found. The capability to intervene at an earlier stage is therefore of very high importance.
A timeline for the terrorist preparations could be described as follows:
- Planning and financing, possibly including theft or robbery
- Obtaining equipment and material
- Preparation and production
- Transportation
- Execution of the attack
In the planning and financing phase, terrorist plots may mainly be revealed through conventional intelligence gathering methods, e.g. surveillance of known terrorist groups, informants, monitoring of money transfers or intercepting the group in a robbery. In many cases, necessary household chemicals or chemicals readily available in hobby stores can be obtained at low cost and little effort. Monitoring trade of restricted materials may be one possible way to reveal terrorist plots at this stage if large amounts of material are purchased or stolen.
The preparation and production phase is by necessity quite long. Unless the bomb-makers are quite professional they will need to test and practice the procedures and in many cases the chemical manufacturing process in itself takes considerable time. Detection in the production phase has several distinct advantages compared to detection in the transportation phase:
- There is more time available for detection;
- Some detected substances may be present in larger quantities;
- If a threat is detected, there is time to take further actions.
Detection during transportation is much more disadvantageous than detection during production. First of all, the detection itself is more difficult and the time frame for intervention activities is also much less as compared to the production phase. Furthermore, even in case of successful detection in the transportation phase, the interruption of the attack is not without risk. If the person carrying the bomb is a suicide bomber, the bomb may be detonated with severe consequences for law enforcement personnel as well as innocent bystanders. Detection during the production phase gives more time for law enforcement to act without risk for themselves or others.
When the plot has reached the last phase - the attack itself - it will be too late and severe damages and casualties will be a fact.
The first large-scale use of IEDs was probably during World War II and ever since IEDs have been used both as an unconventional warfare in military theatres of operation such as in Iraq and Afghanistan and in terrorist attacks such as the Bali (2002), Madrid (2004) or London bombings (2005). IEDs can be manufactured from military and commercially based explosives or home-made explosives and often a combination of these are used in terrorist attacks. The bomb factory may be in an apartment in a large city such as those used in the London Underground attacks and consequently very difficult to both discover and localise.
The terrorist attack executed on the London Underground July 2005 is a horrible example of the result from the manufacture of home-made explosives. The manufacturing of the bombs for the London Underground were performed in the neighbourhood of Leeds. In contrast to the bombs used on July 7 resulting in many victims and injured people, the bombs used on July 21 all failed giving much material for forensic analyses. The bombs were based on a home-made peroxide system where ordinary commercially available chemicals were bought and further modified in a normal kitchen. Many starting chemicals used for making home-made explosives are volatile. Volatile chemicals are possible to detect in the gaseous phase and could be emitted through open windows and kitchen fan systems. The dispersion to the surrounding air is a function of parameters such as weather conditions and city architecture.
An increasing threat to cities is improvised explosive devices made from firework materials. The United Kingdom has a specific task force set up to deal with this issue. It is evident that there exists an urgent need for new tools in the extensive work of preventing future terrorist attacks.
The LOTUS project started 2009 and ends by the finalisation of 2011. The objective of the LOTUS project was to create a system by which illicit production of explosives and drugs can be detected during the production stage. The detectors may be placed at fixed positions although most detectors should be mobile. The size of today's detectors makes placement suitable in vehicles such as a police or other law enforcement vehicles. These distributed detectors continuously sample air while its carrier performs its daily work. The driver of the carrying vehicle does not need to operate the detector and in fact, should not even know if any threats are detected. When a suspicious substance is detected in elevated amounts, information about the type, location, amount and time is registered and sent to a data collection and evaluations centre for analysis. Several indications in the same area will trigger an alert, enabling law enforcement agencies to further investigate and respond. The LOTUS system should be regarded as a technical tool for intelligence gathering, rather than a detection system, meaning that the information obtained should be combined and confirmed with information from other sources to make the most accurate assessment of the situation. LOTUS is a project where a system concept will be developed and tested. For this purpose three substances have been selected for the research and these corresponds to relevant chemicals used when manufacturing drugs or explosives. The substances were chosen from interaction with European end users.
The LOTUS system consists of a number of interconnected and interacting systems. The sensor resides in the roof-top box of a car and communicates through the mobile phone network and the internet to the LOTUS server where the sensor data are sent to and data are analyzed at and further sent and presented on the client workstations containing the LOTUS presentation system. Apart from the sensor, the roof-top box also contains a power supply and an inter-connection unit called router. The router is the central computer in the roof-top box and connects to the sensor for sensor readings, to GPS (Global Positioning System) for geographical location and to the GSM (Global System for Mobile Communications) network for data communication with the server.
The LOTUS client (a presentation system on a computer) showing the LOTUS sensor data on a map can be situated anywhere in the world while the LOTUS sensors is measuring in a city on a completely other place in the world. The only requirement is the access to the GSM and GPS systems but the LOTUS sensors can also store some data if the network is lost and later send this stored data when the network has been found again.
In addition to the hardware components of the LOTUS system some other very important features can be recognized. The transfer of data between the sensor/router to LOTUS server and further to the LOTUS client/presentation software are encrypted in order to achieve a secure and tamper proof communication since the data are security sensitive. In addition, a central evaluation system is positioned at the LOTUS central system in order to minimise false alarms. This involves features named Alarm Confidence Evaluation Functions (ACEF) and Data Fusion (Emitter Location) algorithms.
In the test of the LOTUS system concept, a total of four sensors will be tested. Two sensors are based on Ion Mobility Spectrometers (IMS), one sensor is a Differential Mobility Analyser and one sensor is based on IR absorption spectroscopy using a quantum cascade laser system.
By using an ammonia dopant in one IMS sensor, gas traces from ammonia suppress substances with a low proton affinity and this can reduce the false alarm rate by the increased selectivity and sensitivity. The disadvantage is that some of the interesting substances only can be detected without a dopant. Therefore, two different libraries were developed and tested. One library can be used for the work with ammonia dopant and a second library was developed for the work without dopant or in other words based on the second LOTUS IMS based on water chemistry. One of the IMS sensors is consequently tuned specifically for precursors for drug production and the other IMS and also the DMA sensor is used more generally for a number of other compounds .
The sensors have been tested in the field. For the IMS sensors and the IR absorption sensor these tests were carried out during the summer when the average temperature in the zone was 15-25 degrees Celsius. The tests were carried out at FOIa?s (Sweden) testing area.
The purpose of the trials was to investigate:
- If the IMS can detect drug production starting chemicals in the field;
- The greatest distance at which starting chemicals can be detected by the IMS;
- How much starting chemicals that needs to be discharged for the detection at the different distances?
- Impact of the wind direction on the measurements.
The field trials clearly indicate that the IMS can detect starting chemicals for drugs in the field. The IMS detected these up to a distance of 45 m at a specific release rate/amount and wind conditions. It could be possible to detect at longer distances but it is wind dependant, at greater distances it is essential that the detection point corresponds to the wind direction. There is a clear relationship between the distance of detection and the amount of starting chemical being discharged. The wind direction affects the detection ability.
The IMS was integrated together with the router in a roof top box of a car.
The outside air flows through the tube going right through the roof-top box. The gas inlet of the IMS is connected to this tube by a short distance tube. Solenoid valves can close the gas inlet and the gas outlet if the system is not used in order to avoid contamination of the IMS. The resulting data are transmitted from the IMS via the router to the LOTUS central system.
The roof-top box with the IMS system was mounted on a vehicle. A test was conducted in order to show if the system was able to detect a substance which is emitted nearby the road. After that the car with the IMS system passed the exhausting source, the starting chemicals for drug production was detected.
In a further test, the resulting data of the IMS were sent with the router to the central server. After the combination of the sensor data with the local coordinates the screen with a map showed where the detection of methylamine took place.
It was shown that the chain worked beginning from the detection of the substance by the IMS sensor positioned in the roof-top box via the data transfer to the router and further to the central server until the combination with the coordinates to a display of the detection point in a map on the LOTUS presentation software. This means the operator can evaluate the data remotely on the map/screen anywhere in the world while the sensors are moving around in a city monitoring threat substances. The only requirement is that GSM coverage is needed the most of the time. In the case of insufficient coverage, the system is storing data and sends it as soon as coverage is obtained again.
In similar manners, productions of home-made explosives have been performed in the LOTUS research work. The results are very promising for a particular chosen scenario. Most of the information from these experiments is classified.
In summary, the research work in the LOTUS project encompass development of softwares and hardwares for sending and encrypting data containing position, time and sensor data over the mobile phone network. Real manufacturing of home-made explosives have been performed and the disperions patterns from the indoor production to the outdoor environment have been investigated by conventional wet-chemical analysis techniques as well as using chosen sensors in the system. Dispersions modelings of the relased chemicals in the outdoor environment have been performed. In addition to this, computer algorithms for evaluating sensor data with respect to reducing false alarms and backtracking sensor data to potential release sources (bomb factories) have been developed. The latter part is complex and requires further development.
Manufacturing of home-made explosives and drugs have been performed in a simple kitchen of a mobile trailer in the field. The threat substances in vapor phase from the manufacturing are released to the outdoor environment in concentrations that is possible to detect using the selected sensors of the LOTUS system. This means that substances due to an illicit manufacture of some explosives or precursors to explosives and drugs will be vented out via openings such as kitchen fan exhausts providing the possibility for detection during the preparation phase of terrorist attacks.
The proposed LOTUS system concept, using mobile automatic sensors, data transfer /location via GSM/GPS for on-line detection of illicit production of explosive or precursors to explosives and drugs is a viable approach and is in accordance with historical and today's illicit bomb manufacturing.
If a LOTUS system had been existing 2005 and had been put into operation before the London bombings, the Leeds bomb factory could have been discovered.
The LOTUS system has been found to work in one particular chosen scenario. For other scenarios further research and developments especially concerning sensors are needed.
Project Results:
MAIN S/T RESULTS/FOREGROUND WP100
In this work package, results from field experiments with respect to dispersion measurement studies related to the manufacturing of drugs and explosives have been carried out. These have provided data and knowledge needed for understanding the scenario requirements for the LOTUS project. In addition to the experimental studies performed, end user workshops have been arranged and held two times and the gathered information from the end user meetings have been valuable information on the needs and therefore contributed to the chosen approach for the LOTUS system.
One end user workshops held at FOI facilities in Kista, November 2008, and one workshop held at TNO facilities in Rijswijk, June 2010 have been accomplished. The interactions with end users are very important for understanding many perspectives in a possible deployment of a LOTUS system type in an urban environment. Parameters such as operative and technical possibilities and limitations, legal considerations, public acceptance and cost related issues need to be discussed and addressed. Even though this project aims to develop only a research prototype, awareness on these important parts needs to be brought up so the research performed will take this into consideration which will make the continuation of the technical development easier and focused in the right direction from the beginning.
Some of the more specific objectives of this work package were to perform the dispersion measurement experiments as realistic as possible i.e. in the way an illicit production could be performed in a clandestine laboratory. The experiments are important for understanding the concentration of threat substances in vapour phase at various distances outside an improvised laboratory. This knowledge is crucial for assessing the scenario for the real detection of threat substances using the chosen sensors in the LOTUS project.
The following deliverables in this work package for LOTUS have been reported and are very relevant. They can be accessed on a right to know and need to know basis.
D100.1 Report from end user workshop, Preliminary threat substance list.
D100.2 Threat substance identification, user requirement specification, preliminary scenario requirements specification and preliminary dispersion measurements.
D100.3 Report on dispersion measurements and modelling.
D100.3a Dispersion measurements.
D100.3b Dispersion modelling.
D100.4 Scenario requirements specification.
The results from the field experiments with respect to dispersion measurement studies related to the manufacturing of drugs and home-made explosives are in majority classified. However, some of the result and experimental approach can be given.
The field experiments were conducted in a mobile trailer situated at the FOI all fenced-in test site (GrindsjAn) 40 km south of Stockholm. Inside the trailer, a simple kitchen was installed including a stove, kitchen fan and normal kitchen storage cupboards. An adjacent room used for storing data, safety equipment and protective clothing etc were also present in the trailer. The outer dimensions of the trailer were approximate 7 m x 2.4 m x 2.1 m (length x width x height).
During the drugs or home-made explosives manufacturing, threat substances in vapour phase were sampled indoors of the trailer and outdoors from the trailer at various distances from the release source. The kitchen fan was running during the experiments and windows were also open during some parts of the experiments. The sampling positions for collecting vapours of the target substances (explosives and precursors to explosives and drugs) were tried to be positioned downwind with respect from the position of the kitchen fan exhaust.
As the results of the dispersion measurements are sensitive and partly classified only some information can be given here and only in general terms. First of all, background measurements show very low concentrations of precursors to home-made explosive as they should. Also, the concentrations measured at 10 meters are higher than those measured at 30 meters, again as expected. The measurements also show that the precursor concentration builds up outside the trailer during the experiment as a consequence of the continued experimental work. The most important result is that there are large enough amounts outside the trailer that it is a reasonable assumption that the detectors according to the LOTUS concept can actually detect it.
During the manufacture of drugs, some starting chemicals was detected at 10, 20 and 30 m, however for some chemicals the concentration interval found were in the range to the levels found in the background measurements but for a few other compounds larger concentrations could be detected with variations being due to the distance from the release source.
Disperion modellings have been performed taking the real experimental data into consideration. From a practical perspective it is very important to take the effects of surroundings into account. In the experimental setup the trailer is located on grassland surrounded by wooded areas scattered around it at a distance varying between 50 and several hundreds of meters. One of the purposes of the experiment is to estimate the source term of chemicals that are released during the production of illicit explosives, as well as the dispersion and concentration levels reached at a certain distance from the production facility. In a later stage, these results must be scaled to an urban environment to predict concentration levels in a city.
The main parameter that influences the dispersion of chemicals is the turbulence in the atmosphere. This is a very complex phenomenon which is dependent upon many interrelated parameters like wind speed, temperature, solar radiation, vegetation etc. Around the experimental facility are some features which will influence the turbulence. Due to the variation in woodland and grassland, we also expect additional thermal instability (like thermal updraft), especially during sunshine.
Thermal instability will vary during the day (i.e. morning, afternoon and evening). The width of the grassland area (between the two wood areas) is about 100 meters, which means that the trees will influence the turbulence pattern in the experimental area. The trees adjoining the area will have a channeling effect etc. Integrating these influences into CFD (Computational Fluid Dynamics) is not an easy task as it needs a detailed description of the surroundings and circumstances. Furthermore, many CFD packages still have problems to model the atmospheric boundary level. But perhaps the biggest uncertainty is that, even if we were able to perform a reliable concentration prediction, we cannot predict the circumstances and the weather. This means that we can only predict concentration levels within a certain uncertainty bandwidth of about one order of magnitude. Given this uncertainty, detailed CFD calculations will have hardly any added value compared to a standard Gaussian dispersion model. Therefore, in a first approach, we used a Gaussian dispersion model.
MAIN S/T RESULTS/FOREGROUND WP200
With a view to either preventing production or detecting production and use, the task is to develop novel technologies and innovative approaches for the monitoring of the production, trade, availability and use of chemical components required for the production of certain types of drugs, explosives and agents.
The focus of the LOTUS project is on a detecting productiona?. Since the location of the production site is unknown, which is essential to the antagonists, the main challenge is to devise a feasible approach which has the capability to detect a production site in a reasonable amount of time.
Essentially the LOTUS system works by taking samples in all places where production is likely to occur. Since the size of a place is in the order of tens of meters, this equates to a huge amount of measurements. The workload associated with such a task prohibits anything than an automatic system with mobile sensors.
To cover a large area within a reasonable amount of time a multitude of sensors are needed in parallel. An equal amount of vehicles are required to support sensor mobility. In the long run such a system is only feasible if the sensors may ride piggyback on vehicles that already move in the vicinity of places where production is likely to occur. Thus, sensors need to be mounted on commercial vehicles like buses, taxis, garbage trucks, postal service vehicles, security vehicles etc. that routinely operate where people live and work. Reliable and fast operation under these varying conditions depends heavily on keeping the interaction between vehicle/operator and the sensor system to a minimum.
The LOTUS system addresses these requirements by a communications channel using existing infrastructure for localization and communication. A GPS receiver in the sensor provides the localization information for each measurement instance and the mobile telephone system (GSM) is used for communicating all information to the central server. The communications link between the GSM base station and the central server is of course the internet.
All measurements are made autonomously by the sensor and sent over the communication channel. It is also possible to control the sensor using the same communications channel by issuing commands in the opposite direction.
With potentially sensitive information being communicated over public communication channels it is essential that data is stored and transmitted securely. This is accomplished by utilizing the asymmetric-key cryptosystem for exchanging cryptoa¬graphic keys. The keys are used to secure all data stored in the sensor as well as for securing all transmitted data. In this way all findings of the sensor are intrusion-proof even if the sensor should fall into the hands of the antagonists. The substances identified are numbered, i.e. their chemical names are not available in the sensor software. To further prohibit the antagonists from discovering the presence of a LOTUS system monitoring their neighborhood, all communication is made at random intervals.
All information is stored in a central database. To extract and analyze data and issue alarms two data processing blocks have been developed and integrated into the central server.
The first block ACEF (Alarm Confidence Evaluation Function) evaluates the measurement samples in the database and the second block, Data Fusion, takes metrological and dispersion data into account to calculate the most probable location of the emitter causing the alarm. Both these blocks are governed by a common Project Structure.
The Project concept enables users to set up projects defined by, among others, the geographical area and substances to look for. An arbitrary number of projects may be defined. The project concept enables users to focus on specific tasks and it will also provide confidentiality between different users/projects.
The central server also includes several other functions such as sending alarms by e-mail and the possibility to play and replay the findings of the LOTUS system.
Users communicate with the LOTUS system through client computers whose location is only restricted to an internet connection. Security between the client and central server is maintained by a commercial VPN (Virtual Private Network) solution. The client system LIMS (LOTUS Information Management System) enables direct communication with the sensors, administration of projects, administration of security and display of geographical maps with measurement indications and emitter locations.
The main S&T achievements of WP200 consist not only of developing the system functionality but also of integrating a variety of subsystems and functionalities into a fully functioning tool. Several efforts have been done to reduce the development task, such as:
- The communication protocol between the sensor and the communications router is based on an existing protocol for Bruker sensors.
- The communication protocol between the router and the server is based on the NMEA (National Marine Electronics Association) protocol. This is a communications standard for marine electronics equipment commonly used for communicating with GPS receivers. The router-server protocol used in LOTUS is an extension to the NMEA protocol.
- The security package LSCS (LOTUS Secure Communications System) is based on proven algorithms for exchanging cryptographic keys over public communication channels as well as established secure methods for encrypting data.
- The presentation part of the LIMS (LOTUS Information Management System) is based on an off-the-shelf presentation system and the map-engine as well as the maps is commercial products.
The LOTUS system is a rather complex assembly of subsystems, functionalities and interfaces.
MAIN S/T RESULTS/FOREGROUND WP300
During the production of explosives and drugs higher amounts of precursors are normally present in the air. Because of the illicit substances often have low vapour pressures the direct detection is difficult for gas trace detectors in most cases. But, precursor substances could be used to give an indication of illicit drugs and explosives production sites.
Sensors for the precursor substances were not available on the market in the time of the beginning of the project. Two different technologies were selected at the start point of the project to detect low concentrations of precursors:
- Ion Mobility Spectrometry (IMS)
- Surface Enhanced Raman Scattering (SERS)
The sensors decisions were based in order to meet the following requirements:
- Ability to measure low concentrations of precursors;
- Portable, the sensors should be installed in or on vehicles;
- Fast, the results should be combined with positioning data and implemented in maps with a high local resolution.
Three sensor manufacturers worked at the development and adaptation of the sensors. Bruker adapted an existing Ion Mobility Spectrometer (IMS) for the detection of several precursor substances. The task of Ramem was the development and adaptation of a device based on a special kind of ion mobility spectroscopy, Differential Mobility Analysis (DMA). Portendo switched the primary selected detection method SERS to and IR laser infrared spectroscopy based one. The chance of meeting the requirements was increased by changing the detection technology.
Additionally, a preconcentration method should be developed by Ramem to improve the sample collection and introduction of the sample to one of the detectors that may be required in certain scenarios.
UB should work at data evaluation methods to extract meaningful information from the large amount of sensor data.
Ion Mobility Spectroscopy
Bruker adapted the IMS sensor RAID-S2 to the requirements of the LOTUS project.
The first step was to measure the relevant substances and investigate what configuration of the RAID-S2 is the best to detect a wide range of precursors.
Extensive quantitative measurements were carried out to determine the detection limits and calibration curves for the substances. During the project meeting December 2009 in Leipzig it was decided that the further development should be done with two different configurations of the RAID-S2.
In one configuration, the RAID-S2 works with an internal dopant. The dopant suppresses the influence of interference substances. The disadvantage is that also the detection efficiency for some substances also decreases. In a second configuration, no dopant is used in the RAID-S2. The detection limit of several substances is very low. But, the risk of false alarms increases too.
Special libraries were developed to detect precursors for production sites of drugs or explosives. The air inlet of the device had to be adapted for the use of the instrument in a roof-top box on a car.
The data are evaluated inside the device. Several different possibilities were created to work with the raw and result data. A data communication concept was worked out to use the data in device or analytical investigations and also to test the communication between the sensor and the central station. The device software was adapted to the requirements of the LOTUS data transfer.
Two Bruker RAID-S2 with different configurations were delivered 2010 to the FOI. The devices were used to investigate the dispersion of substances outside in the test area of the FOI in Sweden. In a next step, the devices were integrated into roof-top boxes and several tests were carried out under different metrological conditions. Tests on the test area of the FOI showed that precursors could be detected if a sensor was integrated into a roof-top box on a car.
During the field tests in May/ June 2011, the RAID-S2 devices worked on the top of police and civil cars. The cars went through the cities and the surroundings of Stockholm and Helsinki. The concentration data of gas traces in the air were measured and combined with the position data. The data were sent to the central station by the router at a rate of once a second. With the help of client software the results could displayed in maps with a high local resolution. The field tests showed that the IMS sensors worked under the different conditions in real environment.
Differential Mobility Analysis
RAMEM's goal in the LOTUS project was to develop a high sensitivity, high resolution Ion Mobility Spectrometer on the basis of Differential Mobility Analysis (DMA).
RAMEM's DMA technology has undergone a huge evolution thank to the LOTUS project. At the beginning of the Project the technology was concept-proven but we were lacking an exploitable instrument. Strong R&D efforts have been done to develop different stages of the instrument.
Concerning the ionization, different soft ionization methods have been tested: electrospray, corona and UV. UV has proven to be the most suitable one because of the cleanliness of the spectra, the ease to use, the low cost and the long lifetime. Several UV lamp supplies were tested and finally one was chosen and used due to a combination of performance and cost effectiveness.
Concerning the DMA classification region, the project began with a simple concept that proved a poor performance because it did not take into account the very limited lifetime of ions generated by the UV lamp due to mutual neutralization and collisions. Moreover, the longer the time lapse between the ionization and the detection, the more blurry the ion mobility spectra was and the more difficult the identification of the substance was. Therefore, increasing the sensitivity of the DMA required a decrease in the time lapse between ionization and detection. This translates ultimately into a new design of the DMA in which the UV lamp is very close to the entrance slit of the ion classification volume.
Another requirement of the DMA is the cleanliness of the materials in order to avoid vapours condensing on the internal walls of the instrument. To achieve these goal clean materials such as stainless steel and PEEK were used.
The rest of the instrument was also greatly improved. A heating element was added to avoid internal condensation and to reduce solvation of the ions, which is also a cause of blurry spectra and arcing within the instrument.
Concerning the software of the instrument, a new program was written under LabView environment with three purposes: control of the hardware (measurement of spectra), post-treatment of spectra (graphical toolbox, data treatment) and, in collaboration with the University of Barcelona, the algorithmic to qualitatively and quantitatively characterize the sample. The software proved to have the performance needed for the project, although more development is needed for a real application.
The GPS router was added to the instrument in order to perform the field trials, which were done without any major problems in routes around the downtown and outskirts of Madrid.
All the experience and knowledge acquired in the LOTUS project have been employed to build a new demonstrator which was shown in the LOTUS Final Demonstration in Ossendrecht (The Netherlands) in November 2011. This new demonstrator is more compact, robust and easy to use than the previous versions.
Infrared Spectroscopy
As a result of user requests from the first end-user workshop, held just at the start of the project, one of the intended sensor technologies chosen in LOTUS was re-evaluated. An investigation was made on the basis of the results from the end user and scenario requirements and a laser-based IR absorption was chosen as the technology most likely to meet the user requests.
The LOTUS concept calls for a fast, selective and sensitive sensor. Fast to match the speed of the vehicle, selective otherwise the system would drown in false alarms and sensitive enough to sense the very small amounts available in the air. No doubt several technologies could fulfil one or two of these requirements but recent advances in Infra-Red (IR) lasers made IR absorption the technology most likely to fulfil all three.
Laser-based IR absorption works by measuring the amount of light that has been absorbed during the distance the laser beam has travelled in the medium. The medium here, of course being the ambient air.
To achieve the required sensitivity the laser beam needs to travel around 200m in the ambient air, this was solved by bouncing the beam 200 times between two mirrors spaced 1m apart.
To achieve the required selectivity the laser wavelength has to match an absorption region of the substance in question, which is not also an absorption region for other substances that are likely to be present in the ambient air. One such wavelength was found near 7.9aµm but methane and laughing gas was found at nearby wavelengths requiring a very good resolution in the sensor. This was solved using a brand new type of laser called QCL (Quantum Qascade Laser). The QCL laser is tuneable within a narrow region of wavelengths and by selecting a QCL corresponding to 7.9aµm the measured substance as well as methane and laughing gas could all be separately identified.
Thanks to the high power and rapid tuning speed of the QCL laser the detection speed requirement was also met with this technology. The detection speed enables the sensor to move at normal city speeds (50km/h) according to the determined scenario requirements.
The sensor was mounted in to a roof-box and has been run on a number of occasions, both as part of the project test and demonstration but also as part of media coverage for press and TV.
The technology development in this field is very fast. Although the sensor was designed only two years ago, the latest generation of components would enable twice the performance and a size reduction of ten.
Preconcentration
RAMEM's goal of the LOTUS project was to demonstrate the feasibility of a fast vapor preconcentrator to decrease the limit of detection of the DMA.
As in the case of the DMA, RAMEM's preconcentrator was at the beginning of the project in a very early stage of development. A lab demonstrator was in place early but was not yet proven to work. Several engineering problems related to heat transfer, mechanics and chemical cleanness were encountered.
The preconcentrator needed a fast heating of the carbon bed. The first solution proposed was to heat the air flow before passing through the carbon. This solution proved to work because, given the low heat capacity of air, it decreases a lot in temperature in small duct paths through heat conduction with the walls. Therefore, the first option was to heat up the desorption flow just upstream of the carbon. Even this option did not work, so finally it was decided to heat up the carbon directly. A solution was found with the manufacture of a monolithic duct with a resistor around its walls. This solution proved to work well and was the chosen one.
When dealing with the proof of concept, the problem of cleanliness was encountered. Indeed, a consistent background was found in every spectrum, even in those which were known to be clean enough. A systematic search for contamination was performed, and finally it was determined that the source of contamination was coming from the oil lubricating the pneumatic valves. All the valves underwent a cleaning process compatible with vacuum requirements and reassembled on the preconcentrator.
The proof on the concept was achieved with several substances and reported to LOTUS, obtaining concentration factors in the range 5-50 depending on the substance and on the experimental conditions.
Further improvements of the instrument were related to achieving faster operation and better stability and robustness. One of the main developments was the assembly of an air chiller to cool down the air before starting a new adsorption cycle. This allowed cycle times of 120sec.
The last improvements consisted of changing all pneumatic valves for electrical ones, change all the control electronics for a single PLC module and separating pumps and filters from the module were the carbon beds are, which is the only one accessible by the user.
As a result of all the knowledge acquired in the LOTUS project, RAMEM is building a new prototype of preconcentrator.
DMA Data Processing
University of Barcelona is directly involved in the sub-work package dedicated to the development of the software for processing of data provided by the Differential Mobility Analyzer (DMA) by RAMEM. In order to achieve this issue, UB has been working with spectral data provided by RAMEM S.A. and with own data obtained from a commercial Ion Mobility Spectrometer (IMS) manufactured by Airsense Analytics.
In 2009 some tours were done in Barcelona with a car mounted IMS in order to study the variation of background signals depending on the area or district and the day of the analysis. After analyzing IMS spectra, significant variations in pollution levels were found.
UB started working in deliverable D300.11 the first version of the software (v1.0) for DMA. The first version (v1.0) of the software for DMA was designed and delivered to RAMEM together with the required documentation following the work plan (D300.11 month 20). The software deals with raw data acquired by the DMA and comprises functions for pre-processing (baseline correction, interpolation, filtering, normalization), and functions for instrument calibration/training and real-time prediction (including functions for substance classification and quantification). Since the DMA is expected to provide substance classification and quantification in real-time, two different groups of routines have been created. The first group is for instrument calibration/training so as to prepare the instrument in an initial step (off-line) under known conditions for future real-time operation. The second group of routines is for the prediction of substances (class and concentration) under unknown conditions along real-time (online) field operation. Regarding training and prediction, well-known methods in chemometrics domain have been created and adapted to DMA requirements; for instance, Principal Component Analysis plus Linear Discriminant Analysis (PCA-LDA) for dimensionality reduction leading to substance classification, and Partial Least Squares (PLS) for substance quantification. These methods have been widely tested and have been demonstrated in the literature to be optimal for multivariate data analysis, which is the case for DMA spectra.
Since the acquisition and visualization software for raw DMA spectra has been designed and implemented in LabView 9.0 UB has devoted considerable effort in order to generate a dynamic link library (DLL) of functions capable of running within LabView environment. Therefore, some technical work was needed to be done in order to integrate UB processing routines within the final DMA software, which will allow: spectra acquisition, visualization, data analysis, substance classification and substance quantification. This technical work comprised: MATLAB implementation of the routines, compilation of the software to be adapted to C language standards and generation of the DLL of functions. The first version of the software was successfully tested in the field trials which took place in Madrid in May 2011.
Moreover, UB has been working during 2011 in deliverable D300.12 which has been delivered in month 30. The second version (v2.0) of the software for DMA has been designed and delivered to RAMEM together with the required documentation following the work plan (D300.12 month 30).
The second version of the software is available and is ready for its integration with LabView. UB provides support to RAMEM in this task.
Some improvements have been done with regard to the first version of the software:
For instance, the software is now able to cope with numerical instabilities being more robust.
Moreover, additionally to the functions for substance identification and quantification, some auto-diagnosis functions have been developed.
Some auto-diagnosis functions have been developed in order to endow the software with some smart features:
- The software suggests the number of principal components in PCA-LDA model and the number of latent variables in PLS model. Such a suggestion is based on the recovered variance and the number of samples available to construct the models.
- The software performs cross-validation in order to optimize PLS models.
- Outlier detection. Abnormal samples are identified.
- Possibility to classify samples as NONE
- The software provides a confidence in substance identification based on the distance to each cluster (class/substance)
- The software evaluates the weight of each sample in the construction of the model
A summary of the work was presented as a poster contribution in the LOTUS final meeting in Amsterdam in November 2011.
Summary
During the Demonstration of the sensors in Ossendrecht, The Netherlands in November 2011 it could be shown that all goals for the sensor adaptation and development could be reached. The sensors are able to detect the precursors and work under harsh environmental conditions. During the field tests the sensors worked for a long period of time. The measuring data could be combined with positioning and meteorological data and transmitted to the central station. By the help of the data processing software the large amount of sensor data could be evaluated.
MAIN S/T RESULTS/FOREGROUND WP400
The objective of the work in WP400 is to analyze and control the information flow between sensors and an operation centre. This is realized by the LOTUS Information Management System (LIMS), which has been implemented in the LOTUS system. The progress of the WP400 work during the project was to make a successful integration into the LOTUS System for further verification in the field trails. The progress went as planned and this has laid the foundation for the final demonstration of the LOTUS System.
Saab was assigned to implement the Client System of LIMS (LOTUS Information Management System). The basic functionality required in the System Requirements Specification has been implemented and delivered. The work has included tests and verifications on the Client System during field trials. The progress has gone as planned. The final demonstration has been successfully carried out.
In the project, BNT has been involved in several deliverables; D400.1 D400.3 and D400.4. The majority of work has been put into the Information Analysis Software (D400.3). This includes integrating the Information Analysis System (D400.2) into the Central Server of the LIMS system, storage of the result in a database and making the results available via a web service to the Client systems. Further it includes building a SCIM client, where the operator can monitor the current states of all sensors attached to the system. The MMI (Man-Machine Interface) specification (D400.1) has been updated with a description of this new user interface.
In Work Package 400, UB has been involved in deliverable D400.2B Information Analysis System (month 32). The software (v2.0) and related documentation have been sent to BNT for further work (integration in the LOTUS central system). Some improvements have been done with regard to the main routines (ACEF: Alarm Confidence Evaluation Function and Data Fusion) during 2011. Data formats have been adapted following the guidelines described in D200.3 System design specification.
Improvements have been introduced into the software e.g. a proposal about how to configure the user settings in ACEF and Data Fusion functions. Additionally and a proposal in order to deal with the use of the algorithm has also been put forward and implemented. Therefore, the software is now much more robust against some instabilities.
Furthermore, some simulations were run using the simulator designed during the project and some results were obtained. Simulation results show what happens when the sensor detection limit changes, when wind conditions (speed and direction) change and the system is not aware of this change, and when the size of the area is increased while the number of vehicles is fixed. Simulation results have been described.
The progress of each deliverable is described below
- D400.1 MMI Specification was updated after the field trials. The final version 1.0 has been released and submitted to Commission in month 34.
- D400.2b Information Analysis report v.2 was submitted to the Commission in month 34.
- D400.3 Information analysis software has been integrated into the LOTUS system. Both the Central System and Client System have fulfilled all system requirements in the project. The functions supported have been tested and verified during the field trial and no errors have been reported. Formal software releases have been delivered to the Lotus project for the final demonstration. Therefore the information analysis software D400.3 can be considered completed, the final deliverable was submitted to the Commission in month 26.
The contents of the software consisting of a Central system and Client systems shows that:
- Information presentation in the Client System has been verified and suited for the need of field trails, such as, probability map and background level;
- The Software Performances reported during integration and field trials have been corrected;
- Software User's Manual for the client system including Saab Client and SCIM from BNT have been supplied for end-users;
- All functions in both the Client and Central system have been tested and verified.
Information Management Test Reporting has been submitted and completed. The results conclude:
- It is practically possible to integrate the information flow from sensors to the command centre;
- It is also practically possible to distribute the information to different levels of decision makers over the LOTUS network;
- It has great potentials for practical use of the information presentation.
Further results in this work package includes:
- Allocate meteorological data to sensor measurements;
An interface to the weather service from Swedish SMHI has been implemented and tested during the field trial by receiving weather data from the Stockholm area and allocating them to the sensor data.
- Send alarm notifications via mail.
An alarm notification function has been implemented that sends an email to the operator, if the ACEF calculation results in an alarm for a sensor reading.
- Control of sensors in the Mobile Sensor unit
The LOTUS system has been extended by functions, so the operator can send the required control request to the mobile router from the client system e.g. send a restart command to a sensor.
- Request detailed measurements or log information from Mobile System
The LOTUS system has been extended by functions, so the operator can request a detailed measurement file or a log from a sensor. The files are stored on the Central System.
- Play and replay functions
The Central System has been extended with functions to stores the Background and Emitter probability maps from the data fusion calculations every 15 minutes and provides an interface for the client system to request these maps for replay functions.
Bruker has provided a data set from a field test in Leipzig with the RAID-S2 to UB. The evaluation of the data was performed with modified libraries and the identification thresholds were set to lower values. The purpose of this adjustment was to make it possible to get information about the chemical background of an area.
MAIN S/T RESULTS/FOREGROUND WP500
The objectives of work package 500 were to make a demonstration of the LOTUS integrated system. Towards this objective, the WP was broken in two main tasks: i) Demonstration planning, which included planning of the activities leading up to a demonstration and planning of the demonstration itself, and ii) Capability demonstration which included some of the results from the field trials along with the presentation of the LOTUS system (sensors, vehicle installation and operations' central system) itself. The work carried out in these subtasks, was captured by deliverables D500.1 and D.500.2 respectively.
D500.1 described the activities that were going to take place during demonstration of the LOTUS system. The demonstration plan was based on the recommendations contained in the action Plan for field trials, which included the feedback obtained from the LOTUS Ethics Advisory Board in the project meeting that took place in Athens on 26-27 of January 2011.
D500.2 detailed the activities held during the joint capability demonstration and symposium held on November 2-3, 2011 at the Police Academy in Ossendrecht, The Netherlands.
The LOTUS system was demonstrated by showing the physical mobile sensors and the performance of developed software's and communication system.
The demonstration was held outdoors at the Police academy in Ossendrecht. Three of the LOTUS sensors were shown to the audience. A roof-top box based on the Bruker IMS (Ion mobility spectrometer) detector, a roof-top box based on the IR absorbance spectroscopic technique for detecting precursors to home-made explosives production and a DMA (differential mobility analyser) provided by the Ramem partner. The roof-top box containing the Bruker IMS sensor was driven towards the area for the demo and exposed to a dilute acetone solution by using a simple spray can. The sensor detected this acetone in the air and the sensor data including GPS for positioning was transferred via the mobile phone network to the central LOTUS system (a server positioned in Denmark) and finally downloaded to the improvised command centre at the demonstration area at the Police academy, Ossendrecht. The audience was shown the chain from detection to presentation of data. Moreover, the audience had the possibility to go around and see the different stations such as the three different sensors and the command centre. It was possible to touch and look at the equipment from a close distance, testing the presentation software and discussing with experts from the consortium.
In addition to the demonstration, a symposium was arranged where the LOTUS background, development work and some of the results were presented and discussed by the LOTUS partners. Also, a panel discussion exploring the commercialisation and road-map for future was included. The event was attended by 45 participants originated from nine different member states and organisations in Europe.
One can safely conclude that the three primary objectives for the scope of the demonstration and the LOTUS project itself were met: i) Demonstrate aspects of the detection process and equipment, ii) Establish that the presentation console and the sensors do not have to be physically close to each other, and iii) Illustrate the LOTUS system functionality and effectiveness for detecting selected precursors by integrating the sensors in a networked system using existing technology. Thus the LOTUS system can be used more or less anywhere in the world as a new anti-terrorist tool for law enforcement agencies in certain scenarios.
Potential Impact:
POTENTIAL IMPACT AND DISSEMINATION EXPLOATATION OF LOTUS RESULTS
Going back to the call a?Improving the security of citizensa? there is no doubt that the LOTUS system has the potential to improve the security of citizens. The project has shown that it is possible to locate illicit production of certain chemicals, within a sufficient timeframe. The chemicals are not restricted to those used for making explosives, with respect to the system concept in general, but also include substances that may do harm to the society e.g. drugs, toxic substances and even environmentally harmful substances. For other compounds than those that actually have been tested within the frame of the LOTUS work, primarily development of sensor techniques that are best suited for the detection of chemicals in air in needed scenarios will need to be developed. These sensors would then be easy to incorporate into a LOTUS system for use in this integrated system based on a sensor network.
There are several ways of operating the LOTUS system. It may be used as a continuously monitoring system or as a reactive system used when suspicions, or needs, arise. The central server and client computer may easily be integrated into existing control/supervision centers and the communication channels are already in place.
The short-term potential impact depends on three major factors; building a mobile sensor for daily use without costly maintenance, ensuring that the substance(s) detected matches the substance(s) of interest and a pronounced demand from one or several end-user.
In the long-term, the LOTUS system has a huge potential. Since the infrastructure needed is in place in most countries the main factor driving widespread use relates to the cost, performance and mobility of the sensors.
Sensor mobility depend mainly on size and power consumption, both of these factors are driven by technology development and volume production. Already today the sensors could be mounted inside the car rather than in a roof-top box. Within a couple of years of technology development a sensor may be carried in a back-pack and within a ten-year period it is foreseeable that a sensor may be a medium-cost mobile unit that connects to a mobile phone for communication and location. Performance-wise the sensors will benefit from the current miniaturization of electronics and optoelectronics which will enable detection of multiple substances in one single sensor as well as increased sensing performance.
It is no doubt that the potential applications of the LOTUS system will increase when sensors may be deployed in large numbers and can detect a multitude of substances. These applications will not only include localization of illicit manufacturing of different chemicals, but also detection of environmental hazards and long-term environmental monitoring. The system may also serve to generate statistics of chemicals present in the environment and generate data for the research community. As with many other systems, applications will pop-up once the capability is in place.
From a security perspective a geographically distributed LOTUS system may also aid and facilitate cooperation between nations in security and anti-terrorist matters. The various uses of the system may also facilitate cost-sharing between different end-users.
During the Demonstration of the sensors in Ossendrecht, The Netherlands in November 2011 it could be shown that all goals for the sensor adaptation and development could be reached. The sensors are able to detect the precursors and work under harsh environmental conditions. During the field tests the sensors worked for a long period. The measuring data could be combined with positioning and meteorological data and transmitted to the central station for further evaluations and actions.
If a LOTUS system had been existing 2005 and had been put into operation before the London bombings, the Leeds bomb factory could have been discovered. Consequently, this would have offered police forces the possibility to intervene and catch the terrorists at an early stage of the terrorist activities. This would then have stopped the attack and many lives could have been saved.
Detection of explosives has traditionally been focused on detection when an explosive charge is already ready for use, being transported to or already at the scene of an attack. Preventing terrorist attacks while they are already in motion is extremely difficult. Not only is it difficult to find an explosive on a person or in a bag but one must also face the need to do this in a very short time and then immediately implement measures for countering the attack once a possible suspect or suspicious object has been found. The capability to intervene at an earlier stage is therefore of very high importance. The preparation and production phase is by necessity quite long. Unless the bomb-makers are quite professional they will need to test and practice the procedures and in many cases the chemical manufacturing process in itself takes considerable time. The LOTUS system can be used during primarily the production phase in certain scenarios.
The LOTUS concept is to detect precursors over a wide urban area. The detectors may be placed at fixed positions although most detectors should be mobile. The size of today's detectors makes placement suitable in vehicles such as a police or other law enforcement vehicles. These distributed detectors continuously sample air while its carrier performs its daily work. The driver of the carrying vehicle does not need to operate the detector and in fact, should not even know if any threats are detected. When a suspicious substance is detected in elevated amounts, information about the type, location, amount and time is registered and sent to a data collection and evaluations centre for analysis. Several indications in the same area will trigger an alert, enabling law enforcement agencies to further investigate and respond. The LOTUS system should be regarded as a technical tool for intelligence gathering, rather than a detection system, meaning that the information obtained should be combined and confirmed with information from other sources to make the most accurate assessment of the situation. LOTUS is a project where a system concept has been developed and tested.
The outcome of the research work has produced softwares, hardwares and knowledge on the requirements for the detection of production of home-made explosives and drugs. The research work in the LOTUS project has encompassed e.g. softwares for sending and encrypting data containing position, time and sensor data over the mobile phone network, development of computer algorithms for evaluating sensor data with respect to reducing false alarms and backtracking sensor data to potential release sources (bomb factories) and dispersion modeling calculations in order to understand the dispersion properties of threat substances in vapour phase as released from an indoor laboratory to the nearby outdoor environment. Furthermore, the development of detection techniques and the integration of these into sensors that could make it possible to use them on mobile vehicles in urban areas have made it possible to test the entire LOTUS system concept in real field trials in three capitals in Europe; Stockholm, Finland and Madrid.
It can be concluded that the proposed LOTUS system concept, using mobile automatic sensors, data transfer /location via GSM/GPS for on-line detection of illicit production of explosive or precursors to explosives and drugs is a viable approach and is in accordance with historical and todaya's illicit bomb manufacturing.
If a LOTUS system had been existing 2005 and had been put into operation before the London bombings, the Leeds bomb factory could have been discovered. Consequently, this would have offered police forces the possibility to intervene and catch the terrorists at an early stage of the terrorist activities.
Project Context and Objectives:
One of the largest scientific, technological and deployment challenges in the security domain today is the development of systems for localisation of home-made explosives manufacturing facilities or prevention of the use or enhanced detection possibilities of commercially available precursor chemicals. The improvised explosive devices, that can be prepared in ordinary kitchens, contain explosives that in terrorist attacks give devastating effects. For the law enforcement services there is at present an urgent need for tools that contributes to the early discovery and neutralisation of illicit production of home-made explosives.
The most important and concrete contributions from systems such as LOTUS with respect to the complex task of locating bomb factories or hindering the production are:
- An early warning
- Localisation of the bomb factory
- Automatic system
- Less personnel requirements as compared to normal intelligence
The concept of the surveillance systems for the localisation of IED illicit bomb factories is unique in this area as there today are no commercial systems available.
Detection of explosives has traditionally been focused on detection when an explosive charge is already ready for use, being transported to or already at the scene of an attack. Preventing terrorist attacks while they are already in motion is extremely difficult. Not only is it difficult to find an explosive on a person or in a bag but one must also face the need to do this in a very short time and then immediately implement measures for countering the attack once a possible suspect or suspicious object has been found. The capability to intervene at an earlier stage is therefore of very high importance.
A timeline for the terrorist preparations could be described as follows:
- Planning and financing, possibly including theft or robbery
- Obtaining equipment and material
- Preparation and production
- Transportation
- Execution of the attack
In the planning and financing phase, terrorist plots may mainly be revealed through conventional intelligence gathering methods, e.g. surveillance of known terrorist groups, informants, monitoring of money transfers or intercepting the group in a robbery. In many cases, necessary household chemicals or chemicals readily available in hobby stores can be obtained at low cost and little effort. Monitoring trade of restricted materials may be one possible way to reveal terrorist plots at this stage if large amounts of material are purchased or stolen.
The preparation and production phase is by necessity quite long. Unless the bomb-makers are quite professional they will need to test and practice the procedures and in many cases the chemical manufacturing process in itself takes considerable time. Detection in the production phase has several distinct advantages compared to detection in the transportation phase:
- There is more time available for detection;
- Some detected substances may be present in larger quantities;
- If a threat is detected, there is time to take further actions.
Detection during transportation is much more disadvantageous than detection during production. First of all, the detection itself is more difficult and the time frame for intervention activities is also much less as compared to the production phase. Furthermore, even in case of successful detection in the transportation phase, the interruption of the attack is not without risk. If the person carrying the bomb is a suicide bomber, the bomb may be detonated with severe consequences for law enforcement personnel as well as innocent bystanders. Detection during the production phase gives more time for law enforcement to act without risk for themselves or others.
When the plot has reached the last phase - the attack itself - it will be too late and severe damages and casualties will be a fact.
The first large-scale use of IEDs was probably during World War II and ever since IEDs have been used both as an unconventional warfare in military theatres of operation such as in Iraq and Afghanistan and in terrorist attacks such as the Bali (2002), Madrid (2004) or London bombings (2005). IEDs can be manufactured from military and commercially based explosives or home-made explosives and often a combination of these are used in terrorist attacks. The bomb factory may be in an apartment in a large city such as those used in the London Underground attacks and consequently very difficult to both discover and localise.
The terrorist attack executed on the London Underground July 2005 is a horrible example of the result from the manufacture of home-made explosives. The manufacturing of the bombs for the London Underground were performed in the neighbourhood of Leeds. In contrast to the bombs used on July 7 resulting in many victims and injured people, the bombs used on July 21 all failed giving much material for forensic analyses. The bombs were based on a home-made peroxide system where ordinary commercially available chemicals were bought and further modified in a normal kitchen. Many starting chemicals used for making home-made explosives are volatile. Volatile chemicals are possible to detect in the gaseous phase and could be emitted through open windows and kitchen fan systems. The dispersion to the surrounding air is a function of parameters such as weather conditions and city architecture.
An increasing threat to cities is improvised explosive devices made from firework materials. The United Kingdom has a specific task force set up to deal with this issue. It is evident that there exists an urgent need for new tools in the extensive work of preventing future terrorist attacks.
The LOTUS project started 2009 and ends by the finalisation of 2011. The objective of the LOTUS project was to create a system by which illicit production of explosives and drugs can be detected during the production stage. The detectors may be placed at fixed positions although most detectors should be mobile. The size of today's detectors makes placement suitable in vehicles such as a police or other law enforcement vehicles. These distributed detectors continuously sample air while its carrier performs its daily work. The driver of the carrying vehicle does not need to operate the detector and in fact, should not even know if any threats are detected. When a suspicious substance is detected in elevated amounts, information about the type, location, amount and time is registered and sent to a data collection and evaluations centre for analysis. Several indications in the same area will trigger an alert, enabling law enforcement agencies to further investigate and respond. The LOTUS system should be regarded as a technical tool for intelligence gathering, rather than a detection system, meaning that the information obtained should be combined and confirmed with information from other sources to make the most accurate assessment of the situation. LOTUS is a project where a system concept will be developed and tested. For this purpose three substances have been selected for the research and these corresponds to relevant chemicals used when manufacturing drugs or explosives. The substances were chosen from interaction with European end users.
The LOTUS system consists of a number of interconnected and interacting systems. The sensor resides in the roof-top box of a car and communicates through the mobile phone network and the internet to the LOTUS server where the sensor data are sent to and data are analyzed at and further sent and presented on the client workstations containing the LOTUS presentation system. Apart from the sensor, the roof-top box also contains a power supply and an inter-connection unit called router. The router is the central computer in the roof-top box and connects to the sensor for sensor readings, to GPS (Global Positioning System) for geographical location and to the GSM (Global System for Mobile Communications) network for data communication with the server.
The LOTUS client (a presentation system on a computer) showing the LOTUS sensor data on a map can be situated anywhere in the world while the LOTUS sensors is measuring in a city on a completely other place in the world. The only requirement is the access to the GSM and GPS systems but the LOTUS sensors can also store some data if the network is lost and later send this stored data when the network has been found again.
In addition to the hardware components of the LOTUS system some other very important features can be recognized. The transfer of data between the sensor/router to LOTUS server and further to the LOTUS client/presentation software are encrypted in order to achieve a secure and tamper proof communication since the data are security sensitive. In addition, a central evaluation system is positioned at the LOTUS central system in order to minimise false alarms. This involves features named Alarm Confidence Evaluation Functions (ACEF) and Data Fusion (Emitter Location) algorithms.
In the test of the LOTUS system concept, a total of four sensors will be tested. Two sensors are based on Ion Mobility Spectrometers (IMS), one sensor is a Differential Mobility Analyser and one sensor is based on IR absorption spectroscopy using a quantum cascade laser system.
By using an ammonia dopant in one IMS sensor, gas traces from ammonia suppress substances with a low proton affinity and this can reduce the false alarm rate by the increased selectivity and sensitivity. The disadvantage is that some of the interesting substances only can be detected without a dopant. Therefore, two different libraries were developed and tested. One library can be used for the work with ammonia dopant and a second library was developed for the work without dopant or in other words based on the second LOTUS IMS based on water chemistry. One of the IMS sensors is consequently tuned specifically for precursors for drug production and the other IMS and also the DMA sensor is used more generally for a number of other compounds .
The sensors have been tested in the field. For the IMS sensors and the IR absorption sensor these tests were carried out during the summer when the average temperature in the zone was 15-25 degrees Celsius. The tests were carried out at FOIa?s (Sweden) testing area.
The purpose of the trials was to investigate:
- If the IMS can detect drug production starting chemicals in the field;
- The greatest distance at which starting chemicals can be detected by the IMS;
- How much starting chemicals that needs to be discharged for the detection at the different distances?
- Impact of the wind direction on the measurements.
The field trials clearly indicate that the IMS can detect starting chemicals for drugs in the field. The IMS detected these up to a distance of 45 m at a specific release rate/amount and wind conditions. It could be possible to detect at longer distances but it is wind dependant, at greater distances it is essential that the detection point corresponds to the wind direction. There is a clear relationship between the distance of detection and the amount of starting chemical being discharged. The wind direction affects the detection ability.
The IMS was integrated together with the router in a roof top box of a car.
The outside air flows through the tube going right through the roof-top box. The gas inlet of the IMS is connected to this tube by a short distance tube. Solenoid valves can close the gas inlet and the gas outlet if the system is not used in order to avoid contamination of the IMS. The resulting data are transmitted from the IMS via the router to the LOTUS central system.
The roof-top box with the IMS system was mounted on a vehicle. A test was conducted in order to show if the system was able to detect a substance which is emitted nearby the road. After that the car with the IMS system passed the exhausting source, the starting chemicals for drug production was detected.
In a further test, the resulting data of the IMS were sent with the router to the central server. After the combination of the sensor data with the local coordinates the screen with a map showed where the detection of methylamine took place.
It was shown that the chain worked beginning from the detection of the substance by the IMS sensor positioned in the roof-top box via the data transfer to the router and further to the central server until the combination with the coordinates to a display of the detection point in a map on the LOTUS presentation software. This means the operator can evaluate the data remotely on the map/screen anywhere in the world while the sensors are moving around in a city monitoring threat substances. The only requirement is that GSM coverage is needed the most of the time. In the case of insufficient coverage, the system is storing data and sends it as soon as coverage is obtained again.
In similar manners, productions of home-made explosives have been performed in the LOTUS research work. The results are very promising for a particular chosen scenario. Most of the information from these experiments is classified.
In summary, the research work in the LOTUS project encompass development of softwares and hardwares for sending and encrypting data containing position, time and sensor data over the mobile phone network. Real manufacturing of home-made explosives have been performed and the disperions patterns from the indoor production to the outdoor environment have been investigated by conventional wet-chemical analysis techniques as well as using chosen sensors in the system. Dispersions modelings of the relased chemicals in the outdoor environment have been performed. In addition to this, computer algorithms for evaluating sensor data with respect to reducing false alarms and backtracking sensor data to potential release sources (bomb factories) have been developed. The latter part is complex and requires further development.
Manufacturing of home-made explosives and drugs have been performed in a simple kitchen of a mobile trailer in the field. The threat substances in vapor phase from the manufacturing are released to the outdoor environment in concentrations that is possible to detect using the selected sensors of the LOTUS system. This means that substances due to an illicit manufacture of some explosives or precursors to explosives and drugs will be vented out via openings such as kitchen fan exhausts providing the possibility for detection during the preparation phase of terrorist attacks.
The proposed LOTUS system concept, using mobile automatic sensors, data transfer /location via GSM/GPS for on-line detection of illicit production of explosive or precursors to explosives and drugs is a viable approach and is in accordance with historical and today's illicit bomb manufacturing.
If a LOTUS system had been existing 2005 and had been put into operation before the London bombings, the Leeds bomb factory could have been discovered.
The LOTUS system has been found to work in one particular chosen scenario. For other scenarios further research and developments especially concerning sensors are needed.
Project Results:
MAIN S/T RESULTS/FOREGROUND WP100
In this work package, results from field experiments with respect to dispersion measurement studies related to the manufacturing of drugs and explosives have been carried out. These have provided data and knowledge needed for understanding the scenario requirements for the LOTUS project. In addition to the experimental studies performed, end user workshops have been arranged and held two times and the gathered information from the end user meetings have been valuable information on the needs and therefore contributed to the chosen approach for the LOTUS system.
One end user workshops held at FOI facilities in Kista, November 2008, and one workshop held at TNO facilities in Rijswijk, June 2010 have been accomplished. The interactions with end users are very important for understanding many perspectives in a possible deployment of a LOTUS system type in an urban environment. Parameters such as operative and technical possibilities and limitations, legal considerations, public acceptance and cost related issues need to be discussed and addressed. Even though this project aims to develop only a research prototype, awareness on these important parts needs to be brought up so the research performed will take this into consideration which will make the continuation of the technical development easier and focused in the right direction from the beginning.
Some of the more specific objectives of this work package were to perform the dispersion measurement experiments as realistic as possible i.e. in the way an illicit production could be performed in a clandestine laboratory. The experiments are important for understanding the concentration of threat substances in vapour phase at various distances outside an improvised laboratory. This knowledge is crucial for assessing the scenario for the real detection of threat substances using the chosen sensors in the LOTUS project.
The following deliverables in this work package for LOTUS have been reported and are very relevant. They can be accessed on a right to know and need to know basis.
D100.1 Report from end user workshop, Preliminary threat substance list.
D100.2 Threat substance identification, user requirement specification, preliminary scenario requirements specification and preliminary dispersion measurements.
D100.3 Report on dispersion measurements and modelling.
D100.3a Dispersion measurements.
D100.3b Dispersion modelling.
D100.4 Scenario requirements specification.
The results from the field experiments with respect to dispersion measurement studies related to the manufacturing of drugs and home-made explosives are in majority classified. However, some of the result and experimental approach can be given.
The field experiments were conducted in a mobile trailer situated at the FOI all fenced-in test site (GrindsjAn) 40 km south of Stockholm. Inside the trailer, a simple kitchen was installed including a stove, kitchen fan and normal kitchen storage cupboards. An adjacent room used for storing data, safety equipment and protective clothing etc were also present in the trailer. The outer dimensions of the trailer were approximate 7 m x 2.4 m x 2.1 m (length x width x height).
During the drugs or home-made explosives manufacturing, threat substances in vapour phase were sampled indoors of the trailer and outdoors from the trailer at various distances from the release source. The kitchen fan was running during the experiments and windows were also open during some parts of the experiments. The sampling positions for collecting vapours of the target substances (explosives and precursors to explosives and drugs) were tried to be positioned downwind with respect from the position of the kitchen fan exhaust.
As the results of the dispersion measurements are sensitive and partly classified only some information can be given here and only in general terms. First of all, background measurements show very low concentrations of precursors to home-made explosive as they should. Also, the concentrations measured at 10 meters are higher than those measured at 30 meters, again as expected. The measurements also show that the precursor concentration builds up outside the trailer during the experiment as a consequence of the continued experimental work. The most important result is that there are large enough amounts outside the trailer that it is a reasonable assumption that the detectors according to the LOTUS concept can actually detect it.
During the manufacture of drugs, some starting chemicals was detected at 10, 20 and 30 m, however for some chemicals the concentration interval found were in the range to the levels found in the background measurements but for a few other compounds larger concentrations could be detected with variations being due to the distance from the release source.
Disperion modellings have been performed taking the real experimental data into consideration. From a practical perspective it is very important to take the effects of surroundings into account. In the experimental setup the trailer is located on grassland surrounded by wooded areas scattered around it at a distance varying between 50 and several hundreds of meters. One of the purposes of the experiment is to estimate the source term of chemicals that are released during the production of illicit explosives, as well as the dispersion and concentration levels reached at a certain distance from the production facility. In a later stage, these results must be scaled to an urban environment to predict concentration levels in a city.
The main parameter that influences the dispersion of chemicals is the turbulence in the atmosphere. This is a very complex phenomenon which is dependent upon many interrelated parameters like wind speed, temperature, solar radiation, vegetation etc. Around the experimental facility are some features which will influence the turbulence. Due to the variation in woodland and grassland, we also expect additional thermal instability (like thermal updraft), especially during sunshine.
Thermal instability will vary during the day (i.e. morning, afternoon and evening). The width of the grassland area (between the two wood areas) is about 100 meters, which means that the trees will influence the turbulence pattern in the experimental area. The trees adjoining the area will have a channeling effect etc. Integrating these influences into CFD (Computational Fluid Dynamics) is not an easy task as it needs a detailed description of the surroundings and circumstances. Furthermore, many CFD packages still have problems to model the atmospheric boundary level. But perhaps the biggest uncertainty is that, even if we were able to perform a reliable concentration prediction, we cannot predict the circumstances and the weather. This means that we can only predict concentration levels within a certain uncertainty bandwidth of about one order of magnitude. Given this uncertainty, detailed CFD calculations will have hardly any added value compared to a standard Gaussian dispersion model. Therefore, in a first approach, we used a Gaussian dispersion model.
MAIN S/T RESULTS/FOREGROUND WP200
With a view to either preventing production or detecting production and use, the task is to develop novel technologies and innovative approaches for the monitoring of the production, trade, availability and use of chemical components required for the production of certain types of drugs, explosives and agents.
The focus of the LOTUS project is on a detecting productiona?. Since the location of the production site is unknown, which is essential to the antagonists, the main challenge is to devise a feasible approach which has the capability to detect a production site in a reasonable amount of time.
Essentially the LOTUS system works by taking samples in all places where production is likely to occur. Since the size of a place is in the order of tens of meters, this equates to a huge amount of measurements. The workload associated with such a task prohibits anything than an automatic system with mobile sensors.
To cover a large area within a reasonable amount of time a multitude of sensors are needed in parallel. An equal amount of vehicles are required to support sensor mobility. In the long run such a system is only feasible if the sensors may ride piggyback on vehicles that already move in the vicinity of places where production is likely to occur. Thus, sensors need to be mounted on commercial vehicles like buses, taxis, garbage trucks, postal service vehicles, security vehicles etc. that routinely operate where people live and work. Reliable and fast operation under these varying conditions depends heavily on keeping the interaction between vehicle/operator and the sensor system to a minimum.
The LOTUS system addresses these requirements by a communications channel using existing infrastructure for localization and communication. A GPS receiver in the sensor provides the localization information for each measurement instance and the mobile telephone system (GSM) is used for communicating all information to the central server. The communications link between the GSM base station and the central server is of course the internet.
All measurements are made autonomously by the sensor and sent over the communication channel. It is also possible to control the sensor using the same communications channel by issuing commands in the opposite direction.
With potentially sensitive information being communicated over public communication channels it is essential that data is stored and transmitted securely. This is accomplished by utilizing the asymmetric-key cryptosystem for exchanging cryptoa¬graphic keys. The keys are used to secure all data stored in the sensor as well as for securing all transmitted data. In this way all findings of the sensor are intrusion-proof even if the sensor should fall into the hands of the antagonists. The substances identified are numbered, i.e. their chemical names are not available in the sensor software. To further prohibit the antagonists from discovering the presence of a LOTUS system monitoring their neighborhood, all communication is made at random intervals.
All information is stored in a central database. To extract and analyze data and issue alarms two data processing blocks have been developed and integrated into the central server.
The first block ACEF (Alarm Confidence Evaluation Function) evaluates the measurement samples in the database and the second block, Data Fusion, takes metrological and dispersion data into account to calculate the most probable location of the emitter causing the alarm. Both these blocks are governed by a common Project Structure.
The Project concept enables users to set up projects defined by, among others, the geographical area and substances to look for. An arbitrary number of projects may be defined. The project concept enables users to focus on specific tasks and it will also provide confidentiality between different users/projects.
The central server also includes several other functions such as sending alarms by e-mail and the possibility to play and replay the findings of the LOTUS system.
Users communicate with the LOTUS system through client computers whose location is only restricted to an internet connection. Security between the client and central server is maintained by a commercial VPN (Virtual Private Network) solution. The client system LIMS (LOTUS Information Management System) enables direct communication with the sensors, administration of projects, administration of security and display of geographical maps with measurement indications and emitter locations.
The main S&T achievements of WP200 consist not only of developing the system functionality but also of integrating a variety of subsystems and functionalities into a fully functioning tool. Several efforts have been done to reduce the development task, such as:
- The communication protocol between the sensor and the communications router is based on an existing protocol for Bruker sensors.
- The communication protocol between the router and the server is based on the NMEA (National Marine Electronics Association) protocol. This is a communications standard for marine electronics equipment commonly used for communicating with GPS receivers. The router-server protocol used in LOTUS is an extension to the NMEA protocol.
- The security package LSCS (LOTUS Secure Communications System) is based on proven algorithms for exchanging cryptographic keys over public communication channels as well as established secure methods for encrypting data.
- The presentation part of the LIMS (LOTUS Information Management System) is based on an off-the-shelf presentation system and the map-engine as well as the maps is commercial products.
The LOTUS system is a rather complex assembly of subsystems, functionalities and interfaces.
MAIN S/T RESULTS/FOREGROUND WP300
During the production of explosives and drugs higher amounts of precursors are normally present in the air. Because of the illicit substances often have low vapour pressures the direct detection is difficult for gas trace detectors in most cases. But, precursor substances could be used to give an indication of illicit drugs and explosives production sites.
Sensors for the precursor substances were not available on the market in the time of the beginning of the project. Two different technologies were selected at the start point of the project to detect low concentrations of precursors:
- Ion Mobility Spectrometry (IMS)
- Surface Enhanced Raman Scattering (SERS)
The sensors decisions were based in order to meet the following requirements:
- Ability to measure low concentrations of precursors;
- Portable, the sensors should be installed in or on vehicles;
- Fast, the results should be combined with positioning data and implemented in maps with a high local resolution.
Three sensor manufacturers worked at the development and adaptation of the sensors. Bruker adapted an existing Ion Mobility Spectrometer (IMS) for the detection of several precursor substances. The task of Ramem was the development and adaptation of a device based on a special kind of ion mobility spectroscopy, Differential Mobility Analysis (DMA). Portendo switched the primary selected detection method SERS to and IR laser infrared spectroscopy based one. The chance of meeting the requirements was increased by changing the detection technology.
Additionally, a preconcentration method should be developed by Ramem to improve the sample collection and introduction of the sample to one of the detectors that may be required in certain scenarios.
UB should work at data evaluation methods to extract meaningful information from the large amount of sensor data.
Ion Mobility Spectroscopy
Bruker adapted the IMS sensor RAID-S2 to the requirements of the LOTUS project.
The first step was to measure the relevant substances and investigate what configuration of the RAID-S2 is the best to detect a wide range of precursors.
Extensive quantitative measurements were carried out to determine the detection limits and calibration curves for the substances. During the project meeting December 2009 in Leipzig it was decided that the further development should be done with two different configurations of the RAID-S2.
In one configuration, the RAID-S2 works with an internal dopant. The dopant suppresses the influence of interference substances. The disadvantage is that also the detection efficiency for some substances also decreases. In a second configuration, no dopant is used in the RAID-S2. The detection limit of several substances is very low. But, the risk of false alarms increases too.
Special libraries were developed to detect precursors for production sites of drugs or explosives. The air inlet of the device had to be adapted for the use of the instrument in a roof-top box on a car.
The data are evaluated inside the device. Several different possibilities were created to work with the raw and result data. A data communication concept was worked out to use the data in device or analytical investigations and also to test the communication between the sensor and the central station. The device software was adapted to the requirements of the LOTUS data transfer.
Two Bruker RAID-S2 with different configurations were delivered 2010 to the FOI. The devices were used to investigate the dispersion of substances outside in the test area of the FOI in Sweden. In a next step, the devices were integrated into roof-top boxes and several tests were carried out under different metrological conditions. Tests on the test area of the FOI showed that precursors could be detected if a sensor was integrated into a roof-top box on a car.
During the field tests in May/ June 2011, the RAID-S2 devices worked on the top of police and civil cars. The cars went through the cities and the surroundings of Stockholm and Helsinki. The concentration data of gas traces in the air were measured and combined with the position data. The data were sent to the central station by the router at a rate of once a second. With the help of client software the results could displayed in maps with a high local resolution. The field tests showed that the IMS sensors worked under the different conditions in real environment.
Differential Mobility Analysis
RAMEM's goal in the LOTUS project was to develop a high sensitivity, high resolution Ion Mobility Spectrometer on the basis of Differential Mobility Analysis (DMA).
RAMEM's DMA technology has undergone a huge evolution thank to the LOTUS project. At the beginning of the Project the technology was concept-proven but we were lacking an exploitable instrument. Strong R&D efforts have been done to develop different stages of the instrument.
Concerning the ionization, different soft ionization methods have been tested: electrospray, corona and UV. UV has proven to be the most suitable one because of the cleanliness of the spectra, the ease to use, the low cost and the long lifetime. Several UV lamp supplies were tested and finally one was chosen and used due to a combination of performance and cost effectiveness.
Concerning the DMA classification region, the project began with a simple concept that proved a poor performance because it did not take into account the very limited lifetime of ions generated by the UV lamp due to mutual neutralization and collisions. Moreover, the longer the time lapse between the ionization and the detection, the more blurry the ion mobility spectra was and the more difficult the identification of the substance was. Therefore, increasing the sensitivity of the DMA required a decrease in the time lapse between ionization and detection. This translates ultimately into a new design of the DMA in which the UV lamp is very close to the entrance slit of the ion classification volume.
Another requirement of the DMA is the cleanliness of the materials in order to avoid vapours condensing on the internal walls of the instrument. To achieve these goal clean materials such as stainless steel and PEEK were used.
The rest of the instrument was also greatly improved. A heating element was added to avoid internal condensation and to reduce solvation of the ions, which is also a cause of blurry spectra and arcing within the instrument.
Concerning the software of the instrument, a new program was written under LabView environment with three purposes: control of the hardware (measurement of spectra), post-treatment of spectra (graphical toolbox, data treatment) and, in collaboration with the University of Barcelona, the algorithmic to qualitatively and quantitatively characterize the sample. The software proved to have the performance needed for the project, although more development is needed for a real application.
The GPS router was added to the instrument in order to perform the field trials, which were done without any major problems in routes around the downtown and outskirts of Madrid.
All the experience and knowledge acquired in the LOTUS project have been employed to build a new demonstrator which was shown in the LOTUS Final Demonstration in Ossendrecht (The Netherlands) in November 2011. This new demonstrator is more compact, robust and easy to use than the previous versions.
Infrared Spectroscopy
As a result of user requests from the first end-user workshop, held just at the start of the project, one of the intended sensor technologies chosen in LOTUS was re-evaluated. An investigation was made on the basis of the results from the end user and scenario requirements and a laser-based IR absorption was chosen as the technology most likely to meet the user requests.
The LOTUS concept calls for a fast, selective and sensitive sensor. Fast to match the speed of the vehicle, selective otherwise the system would drown in false alarms and sensitive enough to sense the very small amounts available in the air. No doubt several technologies could fulfil one or two of these requirements but recent advances in Infra-Red (IR) lasers made IR absorption the technology most likely to fulfil all three.
Laser-based IR absorption works by measuring the amount of light that has been absorbed during the distance the laser beam has travelled in the medium. The medium here, of course being the ambient air.
To achieve the required sensitivity the laser beam needs to travel around 200m in the ambient air, this was solved by bouncing the beam 200 times between two mirrors spaced 1m apart.
To achieve the required selectivity the laser wavelength has to match an absorption region of the substance in question, which is not also an absorption region for other substances that are likely to be present in the ambient air. One such wavelength was found near 7.9aµm but methane and laughing gas was found at nearby wavelengths requiring a very good resolution in the sensor. This was solved using a brand new type of laser called QCL (Quantum Qascade Laser). The QCL laser is tuneable within a narrow region of wavelengths and by selecting a QCL corresponding to 7.9aµm the measured substance as well as methane and laughing gas could all be separately identified.
Thanks to the high power and rapid tuning speed of the QCL laser the detection speed requirement was also met with this technology. The detection speed enables the sensor to move at normal city speeds (50km/h) according to the determined scenario requirements.
The sensor was mounted in to a roof-box and has been run on a number of occasions, both as part of the project test and demonstration but also as part of media coverage for press and TV.
The technology development in this field is very fast. Although the sensor was designed only two years ago, the latest generation of components would enable twice the performance and a size reduction of ten.
Preconcentration
RAMEM's goal of the LOTUS project was to demonstrate the feasibility of a fast vapor preconcentrator to decrease the limit of detection of the DMA.
As in the case of the DMA, RAMEM's preconcentrator was at the beginning of the project in a very early stage of development. A lab demonstrator was in place early but was not yet proven to work. Several engineering problems related to heat transfer, mechanics and chemical cleanness were encountered.
The preconcentrator needed a fast heating of the carbon bed. The first solution proposed was to heat the air flow before passing through the carbon. This solution proved to work because, given the low heat capacity of air, it decreases a lot in temperature in small duct paths through heat conduction with the walls. Therefore, the first option was to heat up the desorption flow just upstream of the carbon. Even this option did not work, so finally it was decided to heat up the carbon directly. A solution was found with the manufacture of a monolithic duct with a resistor around its walls. This solution proved to work well and was the chosen one.
When dealing with the proof of concept, the problem of cleanliness was encountered. Indeed, a consistent background was found in every spectrum, even in those which were known to be clean enough. A systematic search for contamination was performed, and finally it was determined that the source of contamination was coming from the oil lubricating the pneumatic valves. All the valves underwent a cleaning process compatible with vacuum requirements and reassembled on the preconcentrator.
The proof on the concept was achieved with several substances and reported to LOTUS, obtaining concentration factors in the range 5-50 depending on the substance and on the experimental conditions.
Further improvements of the instrument were related to achieving faster operation and better stability and robustness. One of the main developments was the assembly of an air chiller to cool down the air before starting a new adsorption cycle. This allowed cycle times of 120sec.
The last improvements consisted of changing all pneumatic valves for electrical ones, change all the control electronics for a single PLC module and separating pumps and filters from the module were the carbon beds are, which is the only one accessible by the user.
As a result of all the knowledge acquired in the LOTUS project, RAMEM is building a new prototype of preconcentrator.
DMA Data Processing
University of Barcelona is directly involved in the sub-work package dedicated to the development of the software for processing of data provided by the Differential Mobility Analyzer (DMA) by RAMEM. In order to achieve this issue, UB has been working with spectral data provided by RAMEM S.A. and with own data obtained from a commercial Ion Mobility Spectrometer (IMS) manufactured by Airsense Analytics.
In 2009 some tours were done in Barcelona with a car mounted IMS in order to study the variation of background signals depending on the area or district and the day of the analysis. After analyzing IMS spectra, significant variations in pollution levels were found.
UB started working in deliverable D300.11 the first version of the software (v1.0) for DMA. The first version (v1.0) of the software for DMA was designed and delivered to RAMEM together with the required documentation following the work plan (D300.11 month 20). The software deals with raw data acquired by the DMA and comprises functions for pre-processing (baseline correction, interpolation, filtering, normalization), and functions for instrument calibration/training and real-time prediction (including functions for substance classification and quantification). Since the DMA is expected to provide substance classification and quantification in real-time, two different groups of routines have been created. The first group is for instrument calibration/training so as to prepare the instrument in an initial step (off-line) under known conditions for future real-time operation. The second group of routines is for the prediction of substances (class and concentration) under unknown conditions along real-time (online) field operation. Regarding training and prediction, well-known methods in chemometrics domain have been created and adapted to DMA requirements; for instance, Principal Component Analysis plus Linear Discriminant Analysis (PCA-LDA) for dimensionality reduction leading to substance classification, and Partial Least Squares (PLS) for substance quantification. These methods have been widely tested and have been demonstrated in the literature to be optimal for multivariate data analysis, which is the case for DMA spectra.
Since the acquisition and visualization software for raw DMA spectra has been designed and implemented in LabView 9.0 UB has devoted considerable effort in order to generate a dynamic link library (DLL) of functions capable of running within LabView environment. Therefore, some technical work was needed to be done in order to integrate UB processing routines within the final DMA software, which will allow: spectra acquisition, visualization, data analysis, substance classification and substance quantification. This technical work comprised: MATLAB implementation of the routines, compilation of the software to be adapted to C language standards and generation of the DLL of functions. The first version of the software was successfully tested in the field trials which took place in Madrid in May 2011.
Moreover, UB has been working during 2011 in deliverable D300.12 which has been delivered in month 30. The second version (v2.0) of the software for DMA has been designed and delivered to RAMEM together with the required documentation following the work plan (D300.12 month 30).
The second version of the software is available and is ready for its integration with LabView. UB provides support to RAMEM in this task.
Some improvements have been done with regard to the first version of the software:
For instance, the software is now able to cope with numerical instabilities being more robust.
Moreover, additionally to the functions for substance identification and quantification, some auto-diagnosis functions have been developed.
Some auto-diagnosis functions have been developed in order to endow the software with some smart features:
- The software suggests the number of principal components in PCA-LDA model and the number of latent variables in PLS model. Such a suggestion is based on the recovered variance and the number of samples available to construct the models.
- The software performs cross-validation in order to optimize PLS models.
- Outlier detection. Abnormal samples are identified.
- Possibility to classify samples as NONE
- The software provides a confidence in substance identification based on the distance to each cluster (class/substance)
- The software evaluates the weight of each sample in the construction of the model
A summary of the work was presented as a poster contribution in the LOTUS final meeting in Amsterdam in November 2011.
Summary
During the Demonstration of the sensors in Ossendrecht, The Netherlands in November 2011 it could be shown that all goals for the sensor adaptation and development could be reached. The sensors are able to detect the precursors and work under harsh environmental conditions. During the field tests the sensors worked for a long period of time. The measuring data could be combined with positioning and meteorological data and transmitted to the central station. By the help of the data processing software the large amount of sensor data could be evaluated.
MAIN S/T RESULTS/FOREGROUND WP400
The objective of the work in WP400 is to analyze and control the information flow between sensors and an operation centre. This is realized by the LOTUS Information Management System (LIMS), which has been implemented in the LOTUS system. The progress of the WP400 work during the project was to make a successful integration into the LOTUS System for further verification in the field trails. The progress went as planned and this has laid the foundation for the final demonstration of the LOTUS System.
Saab was assigned to implement the Client System of LIMS (LOTUS Information Management System). The basic functionality required in the System Requirements Specification has been implemented and delivered. The work has included tests and verifications on the Client System during field trials. The progress has gone as planned. The final demonstration has been successfully carried out.
In the project, BNT has been involved in several deliverables; D400.1 D400.3 and D400.4. The majority of work has been put into the Information Analysis Software (D400.3). This includes integrating the Information Analysis System (D400.2) into the Central Server of the LIMS system, storage of the result in a database and making the results available via a web service to the Client systems. Further it includes building a SCIM client, where the operator can monitor the current states of all sensors attached to the system. The MMI (Man-Machine Interface) specification (D400.1) has been updated with a description of this new user interface.
In Work Package 400, UB has been involved in deliverable D400.2B Information Analysis System (month 32). The software (v2.0) and related documentation have been sent to BNT for further work (integration in the LOTUS central system). Some improvements have been done with regard to the main routines (ACEF: Alarm Confidence Evaluation Function and Data Fusion) during 2011. Data formats have been adapted following the guidelines described in D200.3 System design specification.
Improvements have been introduced into the software e.g. a proposal about how to configure the user settings in ACEF and Data Fusion functions. Additionally and a proposal in order to deal with the use of the algorithm has also been put forward and implemented. Therefore, the software is now much more robust against some instabilities.
Furthermore, some simulations were run using the simulator designed during the project and some results were obtained. Simulation results show what happens when the sensor detection limit changes, when wind conditions (speed and direction) change and the system is not aware of this change, and when the size of the area is increased while the number of vehicles is fixed. Simulation results have been described.
The progress of each deliverable is described below
- D400.1 MMI Specification was updated after the field trials. The final version 1.0 has been released and submitted to Commission in month 34.
- D400.2b Information Analysis report v.2 was submitted to the Commission in month 34.
- D400.3 Information analysis software has been integrated into the LOTUS system. Both the Central System and Client System have fulfilled all system requirements in the project. The functions supported have been tested and verified during the field trial and no errors have been reported. Formal software releases have been delivered to the Lotus project for the final demonstration. Therefore the information analysis software D400.3 can be considered completed, the final deliverable was submitted to the Commission in month 26.
The contents of the software consisting of a Central system and Client systems shows that:
- Information presentation in the Client System has been verified and suited for the need of field trails, such as, probability map and background level;
- The Software Performances reported during integration and field trials have been corrected;
- Software User's Manual for the client system including Saab Client and SCIM from BNT have been supplied for end-users;
- All functions in both the Client and Central system have been tested and verified.
Information Management Test Reporting has been submitted and completed. The results conclude:
- It is practically possible to integrate the information flow from sensors to the command centre;
- It is also practically possible to distribute the information to different levels of decision makers over the LOTUS network;
- It has great potentials for practical use of the information presentation.
Further results in this work package includes:
- Allocate meteorological data to sensor measurements;
An interface to the weather service from Swedish SMHI has been implemented and tested during the field trial by receiving weather data from the Stockholm area and allocating them to the sensor data.
- Send alarm notifications via mail.
An alarm notification function has been implemented that sends an email to the operator, if the ACEF calculation results in an alarm for a sensor reading.
- Control of sensors in the Mobile Sensor unit
The LOTUS system has been extended by functions, so the operator can send the required control request to the mobile router from the client system e.g. send a restart command to a sensor.
- Request detailed measurements or log information from Mobile System
The LOTUS system has been extended by functions, so the operator can request a detailed measurement file or a log from a sensor. The files are stored on the Central System.
- Play and replay functions
The Central System has been extended with functions to stores the Background and Emitter probability maps from the data fusion calculations every 15 minutes and provides an interface for the client system to request these maps for replay functions.
Bruker has provided a data set from a field test in Leipzig with the RAID-S2 to UB. The evaluation of the data was performed with modified libraries and the identification thresholds were set to lower values. The purpose of this adjustment was to make it possible to get information about the chemical background of an area.
MAIN S/T RESULTS/FOREGROUND WP500
The objectives of work package 500 were to make a demonstration of the LOTUS integrated system. Towards this objective, the WP was broken in two main tasks: i) Demonstration planning, which included planning of the activities leading up to a demonstration and planning of the demonstration itself, and ii) Capability demonstration which included some of the results from the field trials along with the presentation of the LOTUS system (sensors, vehicle installation and operations' central system) itself. The work carried out in these subtasks, was captured by deliverables D500.1 and D.500.2 respectively.
D500.1 described the activities that were going to take place during demonstration of the LOTUS system. The demonstration plan was based on the recommendations contained in the action Plan for field trials, which included the feedback obtained from the LOTUS Ethics Advisory Board in the project meeting that took place in Athens on 26-27 of January 2011.
D500.2 detailed the activities held during the joint capability demonstration and symposium held on November 2-3, 2011 at the Police Academy in Ossendrecht, The Netherlands.
The LOTUS system was demonstrated by showing the physical mobile sensors and the performance of developed software's and communication system.
The demonstration was held outdoors at the Police academy in Ossendrecht. Three of the LOTUS sensors were shown to the audience. A roof-top box based on the Bruker IMS (Ion mobility spectrometer) detector, a roof-top box based on the IR absorbance spectroscopic technique for detecting precursors to home-made explosives production and a DMA (differential mobility analyser) provided by the Ramem partner. The roof-top box containing the Bruker IMS sensor was driven towards the area for the demo and exposed to a dilute acetone solution by using a simple spray can. The sensor detected this acetone in the air and the sensor data including GPS for positioning was transferred via the mobile phone network to the central LOTUS system (a server positioned in Denmark) and finally downloaded to the improvised command centre at the demonstration area at the Police academy, Ossendrecht. The audience was shown the chain from detection to presentation of data. Moreover, the audience had the possibility to go around and see the different stations such as the three different sensors and the command centre. It was possible to touch and look at the equipment from a close distance, testing the presentation software and discussing with experts from the consortium.
In addition to the demonstration, a symposium was arranged where the LOTUS background, development work and some of the results were presented and discussed by the LOTUS partners. Also, a panel discussion exploring the commercialisation and road-map for future was included. The event was attended by 45 participants originated from nine different member states and organisations in Europe.
One can safely conclude that the three primary objectives for the scope of the demonstration and the LOTUS project itself were met: i) Demonstrate aspects of the detection process and equipment, ii) Establish that the presentation console and the sensors do not have to be physically close to each other, and iii) Illustrate the LOTUS system functionality and effectiveness for detecting selected precursors by integrating the sensors in a networked system using existing technology. Thus the LOTUS system can be used more or less anywhere in the world as a new anti-terrorist tool for law enforcement agencies in certain scenarios.
Potential Impact:
POTENTIAL IMPACT AND DISSEMINATION EXPLOATATION OF LOTUS RESULTS
Going back to the call a?Improving the security of citizensa? there is no doubt that the LOTUS system has the potential to improve the security of citizens. The project has shown that it is possible to locate illicit production of certain chemicals, within a sufficient timeframe. The chemicals are not restricted to those used for making explosives, with respect to the system concept in general, but also include substances that may do harm to the society e.g. drugs, toxic substances and even environmentally harmful substances. For other compounds than those that actually have been tested within the frame of the LOTUS work, primarily development of sensor techniques that are best suited for the detection of chemicals in air in needed scenarios will need to be developed. These sensors would then be easy to incorporate into a LOTUS system for use in this integrated system based on a sensor network.
There are several ways of operating the LOTUS system. It may be used as a continuously monitoring system or as a reactive system used when suspicions, or needs, arise. The central server and client computer may easily be integrated into existing control/supervision centers and the communication channels are already in place.
The short-term potential impact depends on three major factors; building a mobile sensor for daily use without costly maintenance, ensuring that the substance(s) detected matches the substance(s) of interest and a pronounced demand from one or several end-user.
In the long-term, the LOTUS system has a huge potential. Since the infrastructure needed is in place in most countries the main factor driving widespread use relates to the cost, performance and mobility of the sensors.
Sensor mobility depend mainly on size and power consumption, both of these factors are driven by technology development and volume production. Already today the sensors could be mounted inside the car rather than in a roof-top box. Within a couple of years of technology development a sensor may be carried in a back-pack and within a ten-year period it is foreseeable that a sensor may be a medium-cost mobile unit that connects to a mobile phone for communication and location. Performance-wise the sensors will benefit from the current miniaturization of electronics and optoelectronics which will enable detection of multiple substances in one single sensor as well as increased sensing performance.
It is no doubt that the potential applications of the LOTUS system will increase when sensors may be deployed in large numbers and can detect a multitude of substances. These applications will not only include localization of illicit manufacturing of different chemicals, but also detection of environmental hazards and long-term environmental monitoring. The system may also serve to generate statistics of chemicals present in the environment and generate data for the research community. As with many other systems, applications will pop-up once the capability is in place.
From a security perspective a geographically distributed LOTUS system may also aid and facilitate cooperation between nations in security and anti-terrorist matters. The various uses of the system may also facilitate cost-sharing between different end-users.
During the Demonstration of the sensors in Ossendrecht, The Netherlands in November 2011 it could be shown that all goals for the sensor adaptation and development could be reached. The sensors are able to detect the precursors and work under harsh environmental conditions. During the field tests the sensors worked for a long period. The measuring data could be combined with positioning and meteorological data and transmitted to the central station for further evaluations and actions.
If a LOTUS system had been existing 2005 and had been put into operation before the London bombings, the Leeds bomb factory could have been discovered. Consequently, this would have offered police forces the possibility to intervene and catch the terrorists at an early stage of the terrorist activities. This would then have stopped the attack and many lives could have been saved.