Application of Improved BP Neural Network in Information Fusion Kalman Filter

摘要

Based on improved back propagation (BP) neural network, information fusion state estimation problem for multi-sensor system is considered. Firstly, particle swarm optimization, search dynamic learning rate and additional momentum method are introduced to train the initial weights and thresholds of BP neural network. Then, the improved neural network is used to optimize the estimated value of Kalman filter. Finally, the sate estimators are fused by weighting matrices. A simulation example verifies the effectiveness of the proposed algorithm.