MTACA: A Coverage Control Strategy For Wireless Sensor Network Based On Sensing QoS Oriented

Wireless sensor network (WSN) is an emerging technology focused oninformation gathering. It has grown up with the development of wirelesscommunication, sensor technology, micro-electro mechanical system.WSN has shown great application and research worthiness, particularly inenvironment monitoring.The problem of sensing coverage is the crucialand fundamental issue in the application and research of WSN. It has greatinfluence on the performance of the network.A summarize on the research of WSN coverage control strategy isproposed in the very beginning of this article. On top of that, we give theoptimize methods of coverage control, the influence of the sensing model,the classification of coverage, the restriction issues and the evaluation ofWSN coverage.The sensing model of sensors has significant affect on the coveragecontrol. By summing up the existing research, we propose three models:sensing antitype, 0-1 sensing model and probabilistic sensing model, thelater two of which are discussed in detail. A new concept "CooperativeSensing" is picked up based on probabilistic sensing model.Each node can gather the information of the supervised targets withinits sensing area, while each target can be monitored by several nodes in itsneighborhood, a relationship between the nodes and the targets could beestablished by the sensing process. Starting off from the relationship,MTACA finds out the associative relationship by data mining. Based uponthe coverage requirement of the network and combined with the energyefficiency, we raise a new coverage control strategy: MTACA (MultipleTarget Associate Coverage Algorithm), which holds true with both 0-1 andprobabilistic sensing models. The simulation shows that MTACA haspreferable performance in maintaining coverage, reducing working nodesproportion and extending the network lifetime.In the meantime, the influence on the coverage brought by the nodeslocalization error is studied for the first time. The simulation shows that thelocalization error has great influence on probabilistic-modeled WSN, butlittle on 0-1 modeled WSN.