基于模糊BP神经网络雷达导引头的性能评价

Performance Evaluation of Radar Seeker Based on Improved Fuzzy BP Neural Network

利用模糊系统和BP神经网络的组合模型来进行电子部件的性能评价,组合模型能够兼具二者优点,从理论上比模糊系统性能评价更精准全面。为提高模糊BP神经网络的映射和预测能力,将BP神经网络结构中的动量因子和学习速率两个参数利用误差反馈来调整。从理论上组合模型对电子部件性能评价误差更小,速度更快。通过实例仿真试验表明,改进后的模糊BP神经网络相比较改进前模糊BP神经网络系统,对导引头的性能评价更精准,适用性更强。

The combined model of fuzzy system and BP neural network is used to evaluate the performance of electronic com⁃ponents.The combined model has both advantages and is more accurate and comprehensive than the fuzzy system performance evalu⁃ation theoretically.In order to improve the mapping ability of the fuzzy BP neural network,the momentum factor and learning rate in the BP neural network structure are adjusted by using error feedback.The simulation results show that the improved fuzzy BP neural network has more accurate performance evaluation and stronger applicability,compared with the traditional fuzzy BP neural network.