Adaptive Target Tracking Research In Indoor Complex Environment Based On WSN

Wireless Sensor Networks (WSN) is considered as an emerging area which combines a vatiety of technologies. Target tracking is a significant application of WSN and has been widely used in the smart home, intelligent transportation and environment monitoring, etc.This thesis introduces the principles, advantages and disadvantages of several filtering algorithms. And the specific tracking strategies of WSN and evaluation indicators of target tracking are systematically analyzed. On this basis, an adaptive tracking program is proposed to work in the complex indoor environment which could choose the nodes used in tracking and correct the parameters in the measurement equation of the tracking algorithm as the environment changes.In indoor environment, different areas have different effects on the wireless signal’s transmission as the obstacles are located unevenly, so tracking accuracy can not be guaranteed if only one measurement equation was used. This thesis designs a measurement equation parameters fitting experiment based on the RSSI path loss model in different circumstances by using TI CC2530Development Kit. Based on the measured data in the experiment, least squares fitting method is used to fit the RSSI path loss model in order to give the optimal estimate of the measurement equation parameters which would be stored in the chosen tracking nodes. A adaptive target tracking system based on Gaussian particle filter is designed based on the path loss model, in this system, different tracking nodes and parameters restored in them are selected in different regions, the value of RSSI received by the tracking reference node which blind node send is collected, and then, Gaussian particle filter algorithm is used to achieve the target tracking on the host computer.The results show that, in a range of5m distance, a moving target can be tracked in real time by using adaptive Gaussian Particle Filter which is based on the model of RSSI path loss. The tracking accuracy is maintained in the range of0.5m.