TIRAMISU Final Technology Demonstration at SEDEE-DOVO

By Yann Yvinec, Ph.D., Vinciane Lacroix, Ph.D., and Yvan Baudoin, Ph.D. [ Royal Military Academy of Belgium ] - view pdf

Yvan Baudoin, professor emeritus at the Belgian Royal Military Academy (RMA) and coordinator of the TIRAMISU program, escorts Her Royal Highness Princess Astrid of Belgium, who was honored at the event.
Yvan Baudoin, professor emeritus at the Belgian Royal Military Academy (RMA) and coordinator of the TIRAMISU program, escorts Her Royal Highness Princess Astrid of Belgium, who was honored at the event.
All photos courtesy of Deribaucourt/TIRAMISU.

0n 24 September 2015, and in the presence of Her Royal Highness Princess Astrid of Belgium, the Royal Military Academy of Belgium (RMA) organized a demonstration for mine action tools in Meerdael, Belgium. At the Belgian EOD battalion premises of SEDEE-DOVO, RMA presented technology developed under the TIRAMISU project, which was co-funded by the European Union and coordinated by RMA.1 Designed by 26 different organizations, the tools covered multiple elements of mine action, including survey, operation and validation, information management, mine risk education (MRE), close-in-detection, and personnel protective equipment (PPE). Organizers held inside and outside demonstrations as well as discussions of the tools developed throughout the four-year project.

Several pieces of mechanical equipment with mounted detectors were demonstrated outdoors: a remotely-controlled vehicle with a multi-channel metal detector; an agriculture tractor-based vehicle with a ground penetrating radar and blast resistant wheels, which was operationally tested and certified in Croatia; a small autonomous robot with a rotating metal detector; and a side-looking radar on a vehicle.2,3

Geospatial tools based on open source information and earth observation data were also proposed. In particular visibility and traffic ability analyses, depending on the date of the relative data sets, demonstrated utility for the areas of battle reconstruction, vulnerability estimation or planning demining campaigns. A prioritization tool to determine which areas to address first, based on specific criteria, was also demonstrated with a user case in Cambodia. The use of satellite data, airborne survey and drones to help plan mine action missions and to determine the presence or absence of mine indicators was demonstrated in Croatia. The combination of the generated three-dimensional relief and aerial data has been recognized as a valuable tool for surveyors. A guide on “Geoinformation for demining” is available on the TIRAMISU website, providing the availability of products and services using remote sensing images to assist mine action. Each product or service is described with the data requirements, the required processing, and contact information for companies providing the service.2 The sensors (WorldView-1, WorldView-2, SPOT 2, Landsat 8, etc.) used to generate the products and services, and their characteristics (such as spatial and spectral resolution) are also provided.1

A few of the products demonstrated include the TIRAMISU drone, which provides 3D mapping capabilities, and the TIRAMISAR (in the background on the right), a side-looking imaging sensor with ground penetrating radar.
A few of the products demonstrated include the TIRAMISU drone, which provides 3D mapping capabilities, and the TIRAMISAR (in the background on the right), a side-looking imaging sensor with ground penetrating radar.

Drones were also used in Bosnia and Herzegovina after floods struck the country in 2014. In addition to assisting with search and rescue operations, drones were used to identify locations where floods could have displaced landmines. This was done by building a three-dimensional representation of the terrain.4

To help collect data during surveys, SPINATOR has developed a tablet application called TIRAMISU Information Management System (T-IMS). This application ensures that data are correctly collected with GPS coordinates and easily compatible with other systems such as the Information Management System for Mine Action (IMSMA). It is also linked to the Collaborative ORDnance Data Repository (CORD).

In order to locate all field-based assets, proTime and DIALOGIS designed a set of communication boxes to create a Wi-Fi mesh where GPS coordinates and data can be transferred even in the absence of an internet connection. These boxes can be mounted on mobile equipment, together with a metal detector, in order to gather data and record its position.

Two methods for mine risk education (MRE) were shown. One in particular, designed by Snail Aid, is based on a modular and highly-adaptable theater play broadcast via radio that was evaluated in Algeria and Cambodia.5 The other is a video game from the Institute of Mathematical Machines (IMM) that teaches children MRE safety messages to mitigate the everyday dangers of landmines.

Also presented were new methods by the Spanish National Research Council (CSIC) for training, namely instrumented detectors and prodders for deminers as well as virtual reality applications for operators of remotely-controlled vehicles. An explosives detector vapor was demonstrated by the University of Saint-Andrews (UK) that could be used together with the Remote Explosive Scent Tracing (REST) survey method.

Demonstrators showcased a blast-resistant container designed by the Military Institute of Technical Engineering (WITI) to transport found hazardous items to disposal areas. The container was tested heavily to evaluate its resistance in case on an unwanted explosion. WITI also developed new techniques to dispose of explosives, a method involving the physical destruction of the fuse using hexogen charges.

Most personal protective equipment is currently tested against several consecutive impacts. However, when several impacts occur simultaneously, the equipment can sustain far worse damage. To test equipment against that kind of threat, RMA designed a piece of test equipment with three, adjacent barrels that can shoot three projectiles almost simultaneously at a test object. A film depicting the triple-launcher at work at the RMA ballistic lab was presented to visitors.

Other fundamental research demonstrated at SEDEE-DOVO included the use of honey bees to detect explosive by the University of Zagreb (Croatia) and HCR-CTRO/CROMAC-CTDT. The true challenge does not so much involve training the bees as it does detecting changes in grassy vegetation due to the presence of explosives in the soil.

~Courtesy of Yann Yvinec, Vinciane Lacroix and Yvan Baudoin [ Royal Military Academy ]

MRE & Training Tools

Billy Goat Radio (BGR) is a modular, adaptable system that enables local end-users to create cost-effective MRE campaigns. BGR consists of three elements: a short, radio-broadcasted serial drama; itinerant, live shows; and group discussions. The first use of the tool was a pilot campaign carried out in the Saharawi refugee camps in Algeria. BGR was featured in issue 18.3 of The Journal.5 After completing another campaign in Cambodia in 2014, the tool is now ready to be used and a demo version is available.

~Courtesy of Dr. Emanuela Cepolina

Great Rally is a computer game developed for the TIRAMISU project. Designed for mobile Internet devices as well as devices connected to a local area network, it is a multiplayer game for educational use with teacher-instructor participation, with rally on the backs of electronic turtles through a terrain with mine risks as its content. Pre-tests of the game have been conducted in Poland by children of different ages; full tests were conducted in Croatia.
E-tutor is an e-learning tool for humanitarian demining tasks-management staff, designed to assess staff knowledge and solve site preparation tasks. The tool is composed of two parts, one for an instructor and another for a trainee. There are two parts to the trainee’s tool; the first involves assessing procedural and site preparation, the second involves a map made via the instructor’s tool. After correctly answering questions, trainees try to solve exercises on a map designed by the instructor as a graphical tool for locating necessary facilities.

Training platform is a networked, multi-client solution for training robot operators in humanitarian demining. The simulation process runs on a server, the trainer’s computer, whilst the trainee works on a separate console, which runs the client application. The server is responsible for physics simulation, mission execution and client communication. The client application presents a virtual scenario where the trainee controls a virtual robot. This architecture smoothly runs multiple viewports in simulation without affecting the framerate. Tasks can be designed using multiple types of robots and environments. Simulated environments can be modified freely by adding training equipment (such as road barriers or other graphical objects). Another aspect is task scenario, which is organized with Extensible Markup Language (XML) and Python code. XML script allows trainers to create objects on scene such as robots, training equipment, hitboxes, etc. Python script uses a custom software interface to track the trainer’s progress.

~Courtesy of Andrzej Maslowski and Igor Ostrowski

A few of the products demonstrated include the TIRAMISU drone, which provides 3D mapping capabilities, and the TIRAMISAR (in the background on the right), a side-looking imaging sensor with ground penetrating radar.
The arm of the remotely operated Semi-autonomous Demining Robot Husky-ISR/UC carries a triple coil metal detector, and the robot carries a sensor payload composed of video cameras, a 3D laser range finder, an inertial measuring unit and a GPS receiver.

Close-in Detection Tools

TIRAMI-SAR involves imaging radar at lower microwaves for fast, close-in detection of buried and unburied objects on a larger area within short time (e.g., investigating an area of 50 sq m within a few minutes). This approach allows efficient confirmation of a threat by investigating those regions where detections are present in a secondary step by other sensors.

Different types of matter show different electromagnetic behavior, thus producing different reflectivity. This property can be used to discriminate an object from its surrounding background. For buried objects, the electromagnetic waves have to travel two times through the soil (back and forth), preferably with low attenuation. Since soil moisture predominantly increases the attenuation, especially at higher microwave frequencies, the radar roughly operates at frequencies below a few Gigahertz.

For proper object detection, sufficient spatial resolution is required. Hence the principle of Synthetic Aperture Radar (SAR) is applied. Resolution in ground-range direction is achieved by large signal bandwidth producing narrow pulses in time. In azimuth (cross-range) direction, resolution is achieved by collecting the reflected pulses (radar echoes) over a certain distance called synthetic aperture. All signals are superimposed using known geometry information of the measurement setup and their magnitude and phase information, resulting in a focused image of the reflectivity distribution of the observed area.

SAR for landmine and unexploded ordnance (UXO) detection can be applied by side-looking radar moved on safe ground along the area of interest, which is typically un-safe ground. The ground itself produces radar echoes called clutter, which can confuse proper detection of object echoes. Applying a multi-static observation, i.e., using multiple transmit (TX) and receive (RX) antennas, reduces clutter echoes. In addition, the use of different wave polarization can further reduce clutter and enhance the optimum use of object reflectivity.

Using TIRAMI-SAR, a multitude of optimization and validation experiments have been carried out successfully at DLR facilities in Oberpfaffenhofen, Germany, and at SEDEE-DOVO in Leuven, Belgium. Various advanced processing tools have been developed and applied to enable the proper detection of buried and surface targets from measured data. According to the evaluated results so far, almost all objects, either on the ground surface or buried at a typical depth, could have been clearly detected by the TIRAMI-SAR system. No attempts to discriminate potential threats from false targets, or threats from each other, have been undertaken so far since the primary goal of TIRAMI-SAR was efficient object detection. Discrimination is a complex issue, and was not possible due to the limited duration and resources of the project. However, the TIRAMI-SAR system is considered as valuable tool for further research and optimization of mine/UXO detection using highly advanced radar technology.

~Courtesy of Markus Peichl, Eric Schreiber, Andreas Heinzel, Stephan Dill

Robotics in Demining

The Centre for Automation and Robotics (CAR) CSIC-UPM carried out a live demonstration of an intelligent feedback prodder prototype that has been designed and implemented for training, and a video presentation that demonstrated the remotely-controlled and semi-autonomous operations of a robotic platform for demining applications.

Intelligent prodder for training is an intelligent feedback tool, which provides deminers with information about the amount of force exerted and alerts users when the prodder’s angle is approaching or exceeding a certain limit; the device was designed, implemented and demonstrated by CAR, CSIC-UPM. The training tool consists of a graphical user interface, an instrumented prodder, a data acquisition module and an electronic module for signal conditioning. All basic parts of the instrumented prodder are separable and can be replaced with different extensions in order to obtain different versions of the prodders depending on the user’s needs. The graphical user interface is the principal mechanism through which the instructor supervises the performance of trainees.

Remotely-controlled and semi-autonomous demining vehicle consists of a hexapod robot and a scanning manipulator with 5 degrees of freedom; the robot was designed and manufactured by CAR, CSIC—UPM in Spain. The main objective of this hexapod walking robot is to carry out tasks for localization of antipersonnel (AP) landmines, using an on-board scanning manipulator with a metal detector installed on the tool’s center point. The selective compliance assembly robot arm (SCARA) configuration of the hexapod robot legs allows low-energy consumption when the robot executes gaits on flat terrain or with reduced slope, because its legs are gravitationally decoupled. This legged robot has a mass of about 250 kg (551 lbs), and it can bear a high payload of up to about 300 kg (661 lbs). This capacity for high payload will contribute to reduce the vibrational effects on the manipulator when scanning tasks are performed over the soil. The scanning manipulator is endowed with a set of mini-cameras that enables the 3D mapping of the terrain. With this information, the robotic arm could be controlled to keep the metal detector as closer as possible to the ground, improving the rate of landmines detection.

~Courtesy of Dr. Roemi Fernández, Dr. Héctor Montes and Professor Manuel Armada

Geospatial Tools & Services for General & Non-technical Surveys

Remote sensing imagery and GIS are important assets for mine action, particularly for surveying. Geospatial tools and services are developed to support an Advanced General Survey (AGS) that aims to prioritize areas at the country and regional level. These features will also support Non-technical Surveys (NTS) with the collection and analysis of mine/explosive remnants of war (ERW) contamination data to assess and identify secure hazardous areas (SHA) more precisely.

T-REX is an information management system that includes a customized GIS interface for supporting desk assessments and field surveys, a database for storing and organizing data and metadata, and a collection of map symbols that complement the standardized symbols used in IMSMA. It is based on Free Open Source Software (FOSS), namely QGIS and PostGIS, but any GIS that can interact with PostGIS could also be used as an interface.

T-IMAGE includes services for acquiring remote sensing data and for carrying out pre-processing (e.g., corrections, mosaics, etc.) to provide users with ready-to-use imagery and base maps. Depending on the purpose and size of the area, different platforms can be used: high-resolution satellites, airplanes and remotely piloted aircraft systems (RPAS).

T-MAP uses remote sensing imagery as input and is meant for mapping features of interest according to user requirements via fully automated methods (e.g., bomb craters, trenches, dominant areas), semi-automated methods (e.g., land use, forest edges, roads) and photo-interpretation (e.g., small remnants of fortifications).

T-PRIORITY, alternatively, is specific to AGS and includes an analysis of the socioeconomic vulnerability of the population, support to battleground reconstruction in wide areas with limited information and a web-based decision support system for setting priorities.

T-HS NTS is a hyperspectral aerial survey service for detecting explosive presence indicators. It discriminates between vegetated areas located within minefields from similarly vegetated areas located outside of minefields.

T-SHA aims at reconstructing conflict landscapes in SHAs by combining features extracted from very-high resolution satellite and aerial imagery, and by performing geospatial analyses (e.g., topographic analysis, visibility analysis, optimal routes) to complement existing expert knowledge.

T-EXPL-UXO-DEPOT is an aerial survey service for supporting rapid reaction after the accidental explosion of an ammunition depot. UXO scattered by explosions are mapped using the platform and analysis scheme that best suits the scene characteristics and users’ needs.

Finally, T-AI DSS integrates the tools and services for NTS in an operational system that includes a decision support system for the assessment and identification of SHAs (through exclusion and/or inclusion). Guidelines for users are also provided.

These geospatial tools and services were developed in accordance with user requirements derived from interactions with the Croatian Mine Action Centre (CROMAC), the Cambodian Mine Action Centre (CMAC) and the Cambodian Mine Action and Victim Assistance Authority (CMAA). They are based on actual case studies and are applicable to a wide range of contexts. For more information, please visit the dedicated section of the TIRAMISU website designed for users (http://geospatial.fp7-tiramisu.eu/).

~Courtesy of Sabine Vanhuysse

A few of the products demonstrated include the TIRAMISU drone, which provides 3D mapping capabilities, and the TIRAMISAR (in the background on the right), a side-looking imaging sensor with ground penetrating radar.
Milan Bajic, of the Centre for Testing; Development and Training of the Croatian Mine Action Centre (HCR-CTRO / CROMAC-CTDT), describes the validation tools used in the TIRAMISU project, including evaluation, testing, certification, and operation validation of tools.

Information Management Tools

The TIRAMUS project developed a series of integrated information management tools to support demining actions. The systems combine precision satellite navigation (GPS, GLONASS and GALILEO), with sensors (metal detector, ground penetrating radar, acceleration, gyro, etc.) and communication through mobile data capabilities with web-based location-aware servers, services and geographic information systems (GIS).

The TIRAMISU tool for data Communication and Positioning (TCPbox) is a tool for positioning and data communication at the field level between other tools like sensors and the TRS data repository. It provides precise positioning using Multi-GNSS and a field level AdHoc WiFi network as an information infrastructure. Fitted on the RMA Teodor and connected to the Vallon MDA the TCPbox provides the MDA with positions and compass data. Over the built up AdHoc WiFi communication infrastructure the TCP-Box sends the sensor data and the Teodor track data to the TRS field server.

The Tiramisu Data Repository Service (TRS) is a field server that stores and retrieves any sort of data and files during the demining process. TRS incorporates geographical referencing information data from a wide variety of sources, such as satellite imagery and ground penetrating sensors. It can be used as a rugged device at field-level and also as a hosted system for office level The TRS is connected to the TCPbox WiFi-Mesh-Network and stores the retrieved sensor and positioning data into the TRS Database. The data can be displayed through the DSC tool and also the T-Rex and T-IMS can interact.

The Tiramisu Decision Support Client (DSC) is a mobile spatial management cockpit that integrates different sources of geo-related information and management information in order to support the process for land release planning and decision making in humanitarian demining. Its main goal is to allow non-technical HD expert users to easily explore historical and situational awareness information (i.e. reports, real-time tracking and sensor data) and other area-relevant data stored on the TIRAMISU Repository Service (TRS).

The TIRAMISU Information Management System (T-IMS), is a mobile software application for use in the field. It is built on touch technology and can run on a Windows tablet. T-IMS contains an offline version of the Collaborative ORDnance Repository Database (CORD) for positive identification of UXO.T-IMS supports the use of all well-known map formats, has the ability to layer data, and can exchange information with IMSMA.

~Courtesy of Gerd Waizmann

For more detailed information on the technology featured at SEDEE-DOVO, visit the TIRAMISU website at http://fp7-tiramisu.eu. c



Yann YvinecYann Yvinec, Ph.D., has worked at the Royal Military Academy of Belgium detecting mines, UXO and IEDs for more than 15 years. His experience covers metal detectors and ground-penetrating radars, airborne methods for situation awareness, detection of underwater threats and detector testing and evaluation.

Vinciane LacroixVinciane Lacroix, Ph.D., is a senior researcher at the Royal Military Academy of Belgium, specializing in computer vision. For the past 10 years she has been involved with various humanitarian demining projects, including remote sensing for security applications and image processing for map updating.

Yvan BaudoinYvan Baudoin, Ph.D., served as head of the Department of Mechanical Engineering at the Royal Military Academy of Belgium and as the head of the working group on robotics for humanitarian demining and risky applications in the International Advanced Robotics Programme.

Contact Information

Yann Yvinec
Coordinator of TIRAMISU
Department CISS
Royal Military Academy
Renaissancelaan 30
B-1000 Brussels / Belgium
Tel: +32.2.441.40.42
Email: Yann.yvinec@rma.ac.be
Website: http://www.rma.ac.be/ciss, http://www.fp7-tiramisu.eu/

Vinciane Lacroix
Technical Coordinator of TIRAMISU
Department CISS
Royal Military Academy
Renaissancelaan 30
B-1000 Brussels / Belgium
Tel: +32.2.441.40.42h
Email: Vinciane.lacroix@elec.rma.ac.be
Website: http://www.rma.ac.be/ciss, http://www.fp7-tiramisu.eu/

Yvan Baudoin
Coordinator of TIRAMISU
Department MECA
Royal Military Academy
Renaissancelaan 30
1000 Brussels / Belgium
Email: yvan.baudoin@skynet.be
Website: http://www.rma.ac.be/meca, http://www.fp7-tiramisu.eu/


  1. TIRAMISU: Humanitarian Demining Toolbox. Accessed 8 January 2016. http://fp7-tiramisu.eu.
  2. “Geoinformation for demining.” TIRAMISU: Humanitarian Demining Toolbox. Accessed 5 February 2016. http://bit.ly/1R9NAkU.
  3. Cepolina, Emanuela Elisa and Matteo Zoppi. “Could Local Agricultural Machines Make a Country ‘Impact Free’ by 2010?” Journal of ERW and Mine Action 13.2 (2009). Accessed 6 January 2016. http://bit.ly/1RaT2of.
  4. Bajic, Milan, Tamara Ivelja, Emina Hadzic, Haris Balta, Goran Skelac and Zoran Grujić. “Impact of Flooding on Mine Action in Bosnia and Herzegovina, Croatia, and Serbia.” The Journal of ERW and Mine Action 19.1 (2015). http://bit.ly/22LlUYd.
  5. Scapolla, Luisa and Emanuela Elisa Cepolina. “Billy Goat Radio: MRE in Sahrawi Refugee Camps.” The Journal of ERW and Mine Action 18.3 (2014). Accessed 8 January 2016. http://bit.ly/1ZOfN2Y.