GA-BP神经网络在雷达目标跟踪中的应用研究

目标的运动过程一般是非线性过程,神经网络具有很强的非线性拟合能力,可映射任意复杂的非线性关系,以神经网络为基础的模型能够很好地反应目标的非线性运动趋势。本文在分析传统的BP神经网络的基础上,引入GA遗传算法来优化神经网络的初始权值和阀值,同时将GA-BP神经网络模型运用于对雷达目标的跟踪过程中,并通过仿真验证该模型的精度较高。

Application of GA-BP Neural Network on Radar Target Tracking

Target movement is a nonlinear process. Neural networks have good nonlinear fitting capability, arbitrary complex nonlinear relationship can be mapped by using of neural network. The model based on the neural network can perfectly show nonlinear movement tendency of target. Based on analysis of traditional back-propagation （BP） neural network, a genetic algorithm （GA） is introduced to optimize initial weights and threshold of neural networks. The GA-BP neural network has been applied in radar target tracking process; simulation results verify that high tracking accuracy can be obtained by using of this model.