Agent-based Tracking Method Addressing Target Recognition Errors

Abstract

Target monitoring systems are widely used in various areas such as companies and schools for the prevention of crimes. Such systems require the operators to monitor the information sent from monitoring devices, such as cameras and/or beacon sensors. To reduce the burden on the operators, we have proposed an automatic target tracking system based on mobile agent technologies. The system succeeded in tracking targets when radio frequency identifier (RFID) sensors were used as monitoring devices. However, it sometimes resulted in target recognition errors when cameras were used. In this paper, we propose a method to address target recognition errors. In this method, an agent changes its behavior according to the distance calculated from pictures taken by a camera. We also validated our approach in terms of tracking accuracy through simulation.