Multi-Sensors Data Tracking Fusion Based on a Neural Network Filter

In this paper, we present a multi-sensors data fusion method for target tracking. This approach uses local estimates of the object positions, then the estimates are sent to a central node, where the fusion is done. To achieve a globally optimized performance, these estimates are obtained by the neural network filters using a constant velocity motion model of the target. The coefficients of the filter are estimated by a neural network, then the estimated positions of the target are obtained. Simulation results in three sensors data tracking fusion system are given to show the effectiveness of the proposed algorithm.