DREAM: data fusion and human factors applied to vehicle mounted mine detection

Various sensor technologies are currently being investigated, and, although each solution offers interesting possibilities, it is already clear that no single technology can offer both high detection performance and the level of efficiency required. By associating several types of sensors and using data fusion techniques to cross-cue sensor data, both performance and efficiency can be improved. Such a multisensor system, mounted on a vehicle, will reduce the risk and the cost of demining operations. The latest development of such systems combine sensors. such as ground probing radar, metal detectors and infrared cameras. To be efficient, such high-technology tools have to take into account both the capabilities and experience of users. In order to generate a truly practical concept, the DREAM project has assigned human factors specialists, in close cooperation with sensors and data fusion designers, to manage the technology, to ensure it can be operated by people other than scientists, to ensure it will work in minefields as well as in the test lane and to ensure that errors and accidents are designed-out as much as is possible. This paper outlines the sensor fusion techniques, the human issues addressed, the concerns and problems identified, and indicates a way forward for the development of a practical multisensor vehicle which could be used effectively by non-governmental organisations and commercial concerns.