MENU

Advanced global system to eliminate anti-personnel landmines (apl).

Design, develop and test an innovative integrated system to
neutralise mined areas. This system will enable the
position and kind of mines used to be determined with a 99%
accuracy rate.

Foreword: In the last 25 years more than a million people around the globe - many of them women and children - have been killed or badly injured by antipersonnel mines (APMs). The 1997 Ontario Mine Ban Treaty has now been ratified and on 1st March, 1999 became binding international law, preventing the further use or stockpiling of these horrendous devices. But there are estimated to be upwards of 100 million APMs already in place around the world - and mines can remain active for well over 50 years. Detecting and clearing mines is currently a slow, laborious and dangerous task, using equipment which has not changed much over the last 50 years. Yet as long as mines are known or suspected, productive agricultural land and urban areas alike are effectively unusable. The result is that refugees cannot return and the population is unable to feed itself. The Objective: The ANGEL project aims to create a de-mining system capable of detecting and neutralising or destroying mines buried in minefields. More accurately, the main functional requirement of the system has been stated as: to return mine free area within a given scenario, with an operation performance of at least 2 hectares a day, a reliability of at least 99.6% of mine recognition and clearance, a given operational cost per area and given safety requirements. Types of mines focused on, hereinafter called ANGEL targets, are Antipersonnel Landmines (APLs) (of 4cm. diameter or greater, buried down to 20 cm. or less), and Anti-tank Landmines (ATMs) (buried down to 40 cm. or less), both metallic and non-metallic. APLs and ATMs represent the main goal of the project, although UneXplode Ordnance, unexploded bombs (UXOs), which have also to be cleared in order to deliver safe fields back to civilian use are also addressed. Scenarios addressed initially are those of BOSNIA and HERZEGOVINA (BiH) and the Balkan Region in general. Technical Description: The functionality of the system is divided into four levels. The first level addresses mission planning (i.e. preliminary definition of suspicious areas) using satellite-based Geographic Information Systems (GIS) and databases with information compiled from different sources: Local Authorities, accident reports, etc. The second level performs large area surveys using a multi-sensorial aerial platform in order to identify minefields, that is, area reduction. The third level takes charge of detection and location of each threat in the minefields identified at the previous level, using a multi-sensor ground platform. And finally, the fourth level addresses target neutralisation and destruction using the same previous ground platform but with robotic actuators. Thus the system consists of the following equipment: * a Single Target Location Vehicle (STLV) for the third level function. This is a remotely, tele-operated, All Terrain Vehicle (ATV) platform with: a Ground Penetrating Radar (GPR) and Metal Detector (MD) multi-sensor arrays for single target detection while the vehicle is moving; and biosensor, Ion Mobility Spectrometer (IMS) and X-Ray Spectrometer (XRS) sensors for target confirmation with the vehicle stopped. * A Neutralisation Destruction Vehicle (NSV) for the fourth level function. This is the same ATV type as above, but equipped with remotely-operated robotic arms, capable of placing neutralisation charges, pick-up devices, etc. and clearing large areas using special flails (thus reducing the risks assumed by de-miners). * a Mobile Command and Control (MCC) Centre to coordinate the three introduced subsystems. The information acquired by the different devices will be transmitted to this control room, where advanced software systems will provide Advanced Data Fusion Processing and Computer-Aided Vehicle Tele-Operation using enhanced Human Machine Interfaces (HMI). In spite of the particular architecture presented above, the system analysis carried out during the Definition phase has created a strong interest in the modularity, scaleability and flexibility of the system. This is with the objective of easy integration of new sensor technologies that may be developed in the future. The Human Factor: Given the complexity of the system, trained operators are needed from the beginning in the system's design process, that is, the man in the loop is considered to be part of the system. The ANGEL system works by collecting information and formulating a continuous hypothesis. As soon as the hypothesis reaches given thresholds, fixed at the very beginning to minimize 'false-falses', the system will request a confirmation like a conventional human de-miner). The system has been designed in such a way that any hypothesis provided by the system, either positive or negative, has a supporting logic behind it that can/should be explained on request to the operator. The Automatic Recognition System of the ANGEL system has also focused on the problem of learning convergence. Pure adaptation requires a rich, mode-exciting 'input', which is not the case of the discrete hypothesis framework of the de-mining activity. Other alternatives require the support of a detailed, casuistic, which is again not the case in the de-mining scenario. As a follow-up, techniques for supervised learning convergence were selected, thus selecting 'intelligent systems', such as Expert Systems, adjusting algorithms in order to improve its performance; basically speaking, to zeroing 'false-false' at the same time as keeping the 'false-trues' at reasonable levels. Finally, the ANGEL system assigns this key role to Operators and therefore forces the system to be capable of expressing its 'reasoning' in a way that can be understood by Humans. Finally, the system is being designed to call on humans for support. Two needs should be covered by the Operator: 1. System Learning - Operators should be capable of 'establishing a dialogue' with the Inference Mechanisms of the machine to teach it the right way to decide on a given set of casual circumstances. 2. On Request Tele-Operation - Solving difficulties which have arisen when dealing with the unstructured scenario, such as obstacle avoidance. Savings in system complexity are enormous at the price of low-skilled Human intervention. One should notice that one of the most representative topics of the HMI presented above, is the usage of National Language in the, so-called, Teaching Interface for the 'logic handling' and 'training dialogue' with the Human expert. Project Planning Schedule: The project has been distributed, according to EUREKA's Macao Guidelines, into a 1-year Definition Phase and a 4.5 years Implementation Phase. The Definition Phase, which finished in December 1999 had the following goals: system design and definition and the establishment of the project framework. The 4.5 Implementation Phase has been split into three independent sub-phases each 18 months long: Detailed design, SPAIN Test and BiH Test Phases. The Implementation - Detailed Design Phase (Jan 2000 - June 2001), which is the one currently running, corresponds to the Proof of Concept of the Demonstrator. It aims to evaluate/judge the success opportunities of the technologies to be deployed in the ANGEL project before eventual implementation of the prototype demonstrator. The Implementation - SPAIN Test Phase (Jul 2001 - Dec 2002) aims to build the prototype demonstrator to the Detailed Design provided in the previous phase (first 12 months), and test the prototype demonstrator in the SPAIN Test field (6 months). The Implementation - BiH Phase (Jan 2003 - Jul 2004) aims to improve the prototype demonstrator using the lessons learned during the SPAIN Test (first 12 months) and test the demonstrator in BiH real minefields (6 months). Project Organisation: During the 'Detailed Design' Phase, the partners are organised into three different groups: Air, Ground and Control. A Group Technical Committee follows the activity of the group. One partner in each group will act as Group Coordinator. A Management Committee will follow the activity of the whole project, coordinated by the Programme Manager. Thus, the 'Detailed Design' phase is organised as follows: - ANGEL project partners - Air, Ground and Control Technical Committees - Air, Ground and Control Group Coordinators - Management Committee - Programme Manager. The tasks to be carried out by each partner during this phase are organised on the basis of three main documents: - the System Requirement Document (SRD) which states about 1,000 functional, operational and interface requirements of the main components identified within the ANGEL system. It constitutes the main technical reference of the project. - the Work-Packages Document (WPD), based on the SRD, divides the project development among the partners, and - the Statement of Work (SOW) document states the project structure, the minimal requirements for management, meetings, deliverables, etc. Keywords: landmines, sensors, data fusion.
Acronym: 
ANGEL
Project ID: 
1 889
Start date: 
01-01-2000
Project Duration: 
66months
Project costs: 
32 040 000.00€
Technological Area: 
Sensors/Multisensor Technology, Instrumentation
Market Area: 

Raising the productivity and competitiveness of European businesses through technology. Boosting national economies on the international market, and strengthening the basis for sustainable prosperity and employment.