Unscented Particle Filter with Systematic Resampling Localization Algorithm Based on RSS for Mobile Wireless Sensor Networks

We consider the problem of mobile sensor node localization and propose an unscented particle filter algorithm in wireless sensor networks (WSNs) consisting of mobile nodes and static anchor nodes. Because the received signal strength (RSS) varies obviously, we employ particle filter to decrease the bad effect. We first form the system state model, mobile model, and RSS model, and then apply unscented particle filter and utilize the systematic resample to decrease the degeneration of the particle. We eliminate the uncertain of the RSS in wireless channel and get accurate location of mobile nodes. The predicted position of the mobile node is constrained by its velocity and the measurement value of RSS. We do a lot of simulation to validate the algorithm by assigning different parameters. Simulation results show that the algorithm enhances the localization accuracy of mobile node compared with the standard particle filter algorithm.