An Improved Particle Filter Algorithm Based on Target Tracking in Wireless Sensor Network

Particle filter (PF) is an important way for target tracking in wireless sensor network (WSN). In the paper we proposed an improved particle filter algorithm which outperforms general PF when target suddenly changes movement direction. Our algorithm used estimated direction of motion based on the current measurements to optimize the prediction in PF. It modified the deviation of estimated mean of particle steam that is possibly produced by sudden changes in direction. Simulation results show that the new particle filter algorithm can achieve better tracking performance than other filter algorithms.