Mine detection dogs usually move across minefields, sniffing for TNT and other substances used in mines. This endangers both the dog and the handler, but a remote system, using air vacuumed from an area and passed through a filter, can be remarkably effective, even more effective than some machines that can search an area. A Norwegian scientist, Rune Fjellanger, has developed a system called Remote Explosives Scent Tracing (REST), in which dogs learn to recognized very small amounts of explosives captured in filters. In one circuit of a room, the dog can smell up to 12 filter caught air samples in boxes on a training apparatus consisting of a circular stand with stainless steel arms. A dog alerts to a filter by sitting or lying down. After four to five months of training, dogs can reach a 95% accuracy level or higher and errors can be reduced further by using several dogs on the same samples. If a dog alerts to a filter from a specific area, that area is designated as requiring clearance. Trainers use a reward-based "clicker training," but replace the clicker with a whistle. This system contains significant cost-benefit values over other approaches. A pilot study suggested that this approach be used in Bosnia. R. Fjellanger, E.K. Andersen, and Ian G. McLean, “A Training Program for Filter-Search Mine Detection Dogs,” 15 International Journal of Comparative Psychology 277-286 (2002).
0 comments:
Post a Comment