Wireless Sensor Network Localization Via Compressed Sensing

With the rapid development of wireless sensor network (WSN) technology, WSN basedpositioning system has more wide range of market value and great commercial prospects. Thus, itbecomes a research hot for both academic and industry. Efficient, real-time and precise positioningalgorithm is the key factor of such system applications. Because of simple application and lowcost in received signal strength (Received Signal Strength, RSS) technology, in this thesis, a RSSbased algorithm for WSN localization system is studied in detail.In order to overcome the shortcomings of existing WSN localization algorithm, a two-stageWSN localization is proposed firstly. It took advantages of compressive sensing (CS) technologyand the multilateral measurement. In compressed sensing-coarse positioning stage, the specificlocation of the target in the grid is determined. Meanwhile, in multilateration-fine positioningposition stage, the position of target is estimated finally. It can overcome the limitations of anunknown target position in the grid, reduce the complexity of the algorithm and broaden theapplication scenarios. Secondly, some optimized processing for both two stage are described. In thecoarse positioning stage, using orthogonal matching pursuit compressed sensing reconstructionalgorithm instead of the base track compressed sensing algorithm, it can make the proposedalgorithm keep the same positioning accuracy. However, the time complexity is reduceddramatically. In the fine positioning stage, with the centroid localization algorithm, it can overcomethe shortcomings of the multilateral measurement and improve the target position estimationaccuracy.