Adaptive IMMPF for bearing-only maneuvering target tracking in Wireless Sensor networks
Atiyeh Keshavarz‐Mohammadiyan, Hamid Khaloozadeh · 2016
In this paper, the problem of maneuvering target tracking in a bearing-only Wireless Sensor network (WSN) is considered. To estimate the state variables of the moving target with nonlinear dynamics, Particle Filter (PF) is applied. Moreover, Interacting Multiple Model (IMM) algorithm is used to cope with target maneuvers. To reduce the computational cost of Interacting Multiple Model Particle Filter (IMMPF) algorithm, a novel adaptive sample set size is proposed based on the output estimation error. Besides, a new resampling method is suggested to deal with varying number of samples. The estimation results of the proposed adaptive IMMPF are compared with that of IMMPF with fixed number of samples, in terms of accuracy and computation time.