INTRUSION DETECTION FROM A MOVING PLATFORM

Automatic Target Recognition (ATR) broadly refers to the problems of target detection,acquisition/tracking. The processing is completely autonomous in extracting the necessary information from sensor inputs, in processing, and final decision-making. In addition to the traditional military applications, ATR algorithms can be adopted to endow a mobile sentry robot with the capability to detect intruder generated phenomena. Monocular parallax extracted by tracking objects viewed from a mobile CCD camera is converted to three-dimensional (3-D) position estimates. Simulations have demonstrated that (l) the resolution and accuracy at which a robot can sample its environment using a mobile CCD camera is significantly greater than that achieved using sonar (2) by comparing perceived angular motion of objects in its environment with that expected of Stationary objects, the robot can detect body movement of an intruder while on the move, and (3) 3-D position estimates of objects in the robot's environment can be exploited for autonomous navigation, obstacle avoidance, and map generation. It is proposed that the aforementioned capability be complemented with processing and hardware subsystems which detect different types of intruder generated phenomena as follows: By comparing the 3-D position of objects as determined on the first patrol with that of objects as determined on subsequent patrols, the robot can detect a Stationary intruder or disappearance of possessions. Also, microphones would allow the robot to detect an intruder which it can "hear" not "see."