A nonlinear tracking algorithm of maneuvering target tracking based on MEP-Jerk model

The affect of Jerk model in maneuvering target tracking is discussed in this paper. In order to deal with these problems, an improved algorithm combining the Jerk model with Model Error Predictive filter is proposed. This algorithm can avoid the limitation that the Jerk model assumes the process noise as the Gaussian white noise. Meanwhile, the Model Error Predictive filter estimates the model error online and then amends it. The model error caused as the Jerk model mismatches the real motion is decreased obviously. The Extended Kalman Filter has the problems of complicated calculation, low accuracy and low convergence in estimating the nonlinear target tracking system. To avoid these problems, a new nonlinear filter called Cubature Kalman Filter is adopted. Then the accuracy and stability of the whole target tracking system is enhanced. Numerical simulation results have verified the effectiveness of this algorithm.