基于SVM的来波方向估计方法

提出了一种采用支持向量机(support vector machine, SVM)建立模型实现来波方向估计的新方法。提取已知方向来波信号在天线阵元间感应的相位差作为模型的输入，利用支持向量回归机对复杂函数的逼近能力构建方向估计模型。充分利用了SVM的结构风险最小原则和泛化能力，使得模型对低信噪比和通道误差具有较强的适应能力。通过正弦和余弦函数变换的方法,有效解决了360°~0°转换不连续性引起的逼近误差，提高了估计精度，并给出了算法的具体实现步骤。实验结果验证了该方法的优越性和可行性。

Research on direction of arrival estimation based on SVM

A novel direction of arrival (DOA) estimation method, in which a model is established based on sup-port vector machine (SVM), is presented. The phase differences resulting from known direction signals among the array elements of antennas are extracted as the input of the model, and then the DOA model based on the ability of approximating to complex functions of support vector regression is constructed. The good generalization performance and structural risk minimization theory of SVM is fully utilized so that the model has a good adaptability to the condi-tion of low signal-to-noise ratio and phase errors in the channel. The approximating error is effectively solved by using the transformation of sine and cosine functions, which is caused by the incontinuous transformation from 360° to 0°. The implement process using this algorithm is given in detail. Experiment results prove that the proposed method is good and feasible and has great use in the engineering application.