Confidence in the assessment and use of mine detection systems

In the case of the small, plastic anti-personnel mine detection is generally far from certain-as is evidenced by the current research into developing improved detectors able to cope with many variables. In a properly-designed trial to assess the performance of a candidate detection system, trials would be mounted in which each class of mine was deployed in a representative and defined manner in a representative and specified environment. The detection system would then be swept over the area containing the mines in the recommended manner and the numbers of failed detections and false alarms recorded. The probabilities of detection and false alarm can then be derived from these results. It is the purpose of this paper to establish the sizes of such trials required to obtain assessments that can be believed with confidence.