A Parallel Rough Set Tracking Algorithm for Wireless Sensor Networks

In order to solve speed and accuracy problems of maneuvering target tracking in wireless sensor networks, a Rough Set and neural network Adaptive Interacting Multiple Model (RSAIMM) tracking algorithm was proposed. Based on the establishment of target movement model and rough set neural integration adaptive model, a decision making method which combined rough set theory and neural network was utilized to determine the adjustment of model parameters during target tracking process. Accurate enough system variance was employed to adapt to maneuvering change of the target, thus maintaining rapid response to and high precision tracking of target state. Simulation result showed that over traditional Interacting Multiple Model (IMM) tracking algorithm, tracking accuracy of RSAIMM algorithm improved by 23.15%, and convergence speed of RSAIMM algorithm improved by 21.67%.