改进的BP算法在多传感器数据融合中的应用

采用了基于小波神经网络的BP权值平衡改进算法,构造小波神经网络并训练以改变BP网络权值.根据多传感器特征级数据融合模型,并结合该权值平衡算法,使测量到的数据进行基于特征级的融合,并将该数据融合结果提供给决策级判断,从而得出理想的判定效果.仿真结果表明,该数据融合算法避免了BP权值平衡算法的缺点,不仅提高了学习的速度,而且具有更高的计算精度.

Research on multi-sensor data fusion based on improved BP algorithm

An improved BP(back propagation) weight balance algorithm based on wavelet neural network(WNN) is proposed,in which a WNN is established and trained to change the BP weights.According to the multi-sensor data fusion model at feature level,with the weight balance algorithm,the data fusion of the measured result is carried out at the feature level and the fusion result is provided as a judgment basis of the decision-level.Finally,an ideal determination is obtained.Simulation results indicate that the data fusion algorithm,which avoids the shortcoming of BP weights balance algorithm,not only increases the learning speed but also brings higher precision.