Research On Multi-node Collaborative Target Detection And Tracking In Wireless Sensor Networks

Target detection is one of the most important functions of the wireless sensornetwork, a large number of sensor nodes deployed in the monitored area detect thetarget. The sensor nodes are powered by battery，once energy of the node is below acertain threshold or energy depletion, it will affect the performance of the entirenetwork. This requires us to design the agreement must be considered energy efficientuse of the network, reduce energy consumption of the network, to extend the networklifetime.This paper compares the current domestic and international wireless sensornetworks for target detection and tracking protocol algorithm, and analyzes the variousprotocols of target detection algorithms for different pros and cons. According to thecharacteristics of wireless sensor networks, this paper presents the boundary node basedon energy efficient selection algorithm and multi-node cooperative target detection andtracking protocol. The main research works can be summarized as follows:1. We propose an energy efficient boundary node selection algorithm. Selecting theboundary nodes by the right-hand rule and then filtering the redundant boundary node,which ensures the effectiveness of the selected boundary node. The boundary nodessend messages to the adjacent nodes, and the entry time and entry position are recorded,and target tracking is realized through the node response mechanism. Comparing withother boundary node algorithms, our algorithm has better performances in the numberof boundary nodes, the average number of data packets, the time of target detection andresidual energy.2. We propose multi-node cooperative target detection and tracking algorithm, themonitoring area is divided into a number of cells. Node used sleep schedulingmechanism for target detection to ensure a unit have one node is active. Once the clusterhead node discovery targets will wake up the node of in-cluster to locate and track ofthe targets. Since the nodes are randomly deployed in the monitoring area, which makesthe sensing node in the monitoring area can not be completely covered area, there maybe blind, or the like for the critical node runs out of energy loss lead to the target loss.When the target is lost, the search mechanism will start until find out the target. If thesearch does not find the boundary region to reach the target area, the target is consideredlost. The algorithm in the target detection probability, the target loss rate, energyconsumption has better performance in comparison other algorithms.