Research On Target Location Method Of Wireless Network Based On Compression Sensing

With the rapid development of computer network technology and wireless sensor network technology, the location-based services became the most development potential Internet business. If the location information of the object can’t be accuratly, then smart home, intelligent warehouse and so on based on wireless sensor networks will not be able to achieve its good performance, they can’t achieve the real intelligence. At present, the research on localization are mostly based on wireless nodes. In this paper, localization based on wireless nodes are also be reseached, besides, a new localization methed called device free localization(DFL) will be reseached. The novel compressed sensing (CS) theory can largely reduce the sampling rate, reduces the computational complexity, improving the performance and effect of wireless localization.Firstly, the fundamental theory and key technologies of compressed sensing(CS) and bayesian compressed sensing(BCS) are researched. At the same time, some experimental simulation based reconstruction algorithm are be accomplished.In order to achieve the goal of target localization based wireless nodes effectively, TinyOS system and Telosb nodes are selected as the software and hardware platform. The communication program of the unknown node and beacon node and base station are carried out based on this platform. The curve of different nodes between RSSI and distance were confirmed by experiments in the laboratory and outdoor environment. And also, the effects of human disturbance on the test results were analyzed. Then, the target localization based above experimental data for wireless nodes has been accomplished.To achieve effective DFL, a localisation algorithm based on the CS theory has been proposed. A dynamic statistical model to relate the link measurements with the target’s location and formulated the localisation problem as a sparse signal reconstruction question is presented. And then, the CSBGMP algorithm to realise signal reconstruction and location estimation is proposed. The experimental evaluations confirmed our proposed schemes, and revealed that the CSBGMP algorithm could achieve reasonable performance with few measurements.