基于RBF神经网络的二相码旁瓣抑制

研究了RBF神经网络在二相码雷达旁瓣抑制中的应用,对神经网络的学习算法进行了改进,采用Levenberg-Marquardt（LM）算法优化隐层神经元的中心值和扩展常数,而用最小二乘（LS）法优化隐层至输出层的连接权值。对13位巴克码进行了仿真。仿真结果表明,改进的算法具有极快的收敛速度,可获得60dB以上的输出峰均值比,提高了雷达的探测性能。

Binary-coded Side-lobe Suppression Based on RBF Neural Network

This paper studies the application of radial basis function（RBF）neural network to the side-lobe suppression in binary-coded radar,improves the learning algorithm of neural network,adopts the Levenberg-Marquardt（LM）method to optimize the center value and the extended constant of neuron in the hidden layer,and adopts least-square（LS）method to optimize the connecting weight value from hidden layer to output layer,simulates the 13bit Barker code.The simulation results show that the improved algorithm has faster convergence speed,can obtain over 60dB ratio of output peak value to mean value,and improves the detection performance of radar.