遗传BP网络在雷达装备BIT虚警抑制中的应用研究

Research on Application of GA-BP Neural Networks in Reducing BIT False Alarm of Radar Equipment

摘要:

针对现有故障诊断方法不能有效抑制雷达装备BIT虚警的现象,提出一种遗传算法优化后的BP神经网络技术抑制雷达装备BIT虚警的方法；首先介绍经遗传算法优化后BP神经网络的基本结构和学习算法,再结合雷达装备BIT的特点,以某火控系统雷达发射机作为被故障诊断对象,采用9个具有代表性的雷达发射机故障特征和8个典型故障,以Matlab作为开发工具进行仿真实验；实验结果表明该方法能准确对故障进行定位,有效抑制BIT虚警,提高雷达系统故障诊断能力。

Abstract:

This paper proposes a false alarm reduction method on radar equipment BIT based on BP neural networks optimized by Genetic algorithm in order to eliminate BIT false alarm which is reduced ineffectively by existing fault diagnosis methods. Firstly this paper introduces the principle and corresponding learning algorithm of BP neural networks optimized by Genetic algorithm. Then in view of the characteristics of radar BIT, this paper constructs a fault diagnosis system based on GA-BP neural networks by using Matlab to simulate the process of fault diagnosis with nine respective fault characteristics and eight typical fault examples of Radar transmitter. The diagnosis result indicates that the GA-BP neural networks can identify and locate the fault of sample effectively, and the false alarm of BIT can be reduced, which improves the fault diagnosis ability of radar system.