基于GA-BP神经网络的雷达干扰效能评估方法

Effectiveness evaluation method of radar jamming based on GA-BP neural network

摘要:

针对当前雷达干扰效能评估方法中评估模型较为复杂、参数获取困难、应用价值不大的问题，优化了雷达干扰效能评估变量和指标体系，使其更加贴近实际应用；针对传统雷达干扰效能评估方法中依赖专家打分、人为因素影响较大而普通神经网络预测误差较大的问题，采用遗传算法（genetic algorithm, GA）对误差反传（back propagation, BP）神经网络的初始参数进行全局优化，提出基于GA-BP神经网络的雷达干扰效能评估方法,降低评估系统误差。最后，进行了仿真验证，与普通BP神经网络和支持向量机（support vector machine, SVM）进行了对比分析，并通过调整参数进一步优化了该方法。仿真结果表明，该方法明显优于普通BP神经网络和SVM，具有较好的准确度和稳定性，可为实际应用提供科学依据。

Abstract:

Aiming at the problems in the current radar jamming effectiveness evaluation methods that the evaluation models are complex, the parameters are difficult to obtain and the application value is little, the variable and index system of radar jamming effectiveness evaluation are optimized to be closer to practical applications. Aiming at the problem that expert scoring is relied on and human factors are existed in traditional radar jamming effectiveness evaluation methods, and large prediction errors are existed in ordinary neural networks, a radar jamming effectiveness evaluation method based on GA-BP neural network is proposed, and genetic algorithm(GA) is used to globally optimize the initial parameters of back propagation(BP) neural networks to reduce the evaluation system error. Finally, the method is simulated and verified, compared with ordinary BP neural network and support vector machine(SVM), and further optimized by adjusting the parameters. The simulation results show that the performance of the methodwith good accuracy and stabilityis significantly better than that of ordinary BP neural network and SVM, with good accuracy and stability, which can provide scientific basis for practical application.