关键词: 覆盖质量评估, 多目标, 小波神经网络, 模糊推理, k均值聚类

Abstract: For resolving the problem about coverage quality evaluation of wireless sensor networks (WSNs) with 2 layers, a multiobjective evaluation method of coverage quality based on hybrid algorithm of fuzzy wavelet and clustering is proposed. Firstly, the conception of network unit is proposed, and a model of wireless sensornetworks with 2 layers is established. Secondly, every subobjective is preprocessed at the gathing endnodes. Thirdly, twicealarmed mechanism is introduced, which includes two parts. One part is analysis and fusion subsystem based on fuzzy wavelet neural network. It realizes the first alarm, which checks out the ultrainefficient coverage units. The other part is decision and output subsystem. In this part, multidegree evaluation architecture for network units is proposed, which is based on kmeans clustering algorithm. Meantime, it achieves the second alarm and shows all of inefficient coverage units. The experimental results validate the proposed method from comprehensive evaluation about coverage, consumption, and transmission. This method can determine inefficient coverage units accurately. Also, it conduces to redeployment timely for healthy running WSNs.

Key words: coverage quality evaluation (CQE), multiobjective, wavelet neural network (WNN), fuzzy reasoning, kmeans clustering