Decentralized Target Search in Topology Maps Based on Weighted Least Square Method

摘要

A distributed target search and predictive approach that uses wireless sensor network topology map is presented in this paper. The decision-making and information gathering is distributed, and the target prediction is based on the weighted least square method. The knowledge of target location and moving direction is obtained by time sets recorded by each sensor node. Time set is a collection of time values that describes when the target was detected by a particular node. All the locations are calculated in a topology map, which is an economical alternative to physical map. The results show that the proposed method has been able to search the target, even though the target changes the moving pattern from time to time. In addition, the robot avoids the obstacles and searches for the target in an effective way.