A Comparison between Centerlized and Distributed agent-based Evolutionary Target Tracking System

This paper introduces an agent-based target tracking system. A comparison between two communication schemes among these agents and a central processing unit to provide the target tracking task is also provided. The target tracking is done in two stages. In the first stage, genetic algorithms is implemented to insure optimum coverage of the available agent-sensors over the target area. Sensors are activated based on need according to the tracking requirement of the moving object. To optimize power usage and hence minimize human intervention for maintenance purpose, unneeded sensors are kept in a low-power stand-by state. In the second stage Sensor-agents can communicate among each other through either a centralized or a distributed communication scheme. Results show that distributed system, in which sensor-agents have direct communication among themselves without the need for a central processing unit, proves to be more efficient.