Environmental Monitoring in Wireless Sensor Networks using Structured Matrix Completion

Environmental monitoring is an important application of wireless sensor networks (WSNs). However, due to the limited number of sensors, the global distribution of the sensed physical environmental parameter with a high resolution in the monitoring area cannot be accurately obtained. In this paper, a structured matrix completion based method is proposed to obtain the global distribution of selected environmental parameter with partial sensors. By arranging the data into an enhanced matrix exhibiting Hankel structure, the inherent correlation among data in the monitoring area can be further exploited to improve the accuracy of data estimation. Furthermore, an efficient algorithm based on alternating direction method of multipliers is described to solve the resulting problem. Experimental results demonstrate that the proposed method can estimate the global distribution of the environmental parameter and achieves better estimation accuracy compared with the existing methods.