Target Tracking Based on Particle and Kalman Filter Combined with Neural Network

Wide variations of the maneuvering targets must be track by radar systems. One of the main problems is adaptation of the tracking filter to real behavior of the target. Combining real-time of Kalman filter with robustness of particle filter and neural network, proposes a method for target movement with non-linear motion. The perspective system of adaptive radar data processing for 2D tracking system based on particle filtering is described in this paper. The proposed system has better performance of state estimation (converges to true flight position) then linear Kalman filtering or non-linear adaptive Kalman filtering with Neural Network adaptation. The system can track or estimate object position in environment with interference and noise. The results of simulation show that the method can effectively improve the target tracking process. Final estimated track of object is given by coordinates represented by particles with the highest value of their weights. The modern system of adaptive radar data processing for 2D coordinate system based on particle filtering is designed in this paper.