基于多径参数模型和混合优化的时变水声信道跟踪

Tracking of time-varying underwater acoustic channels based on multipath parameter model and hybrid optimization

针对海洋水声信道的稀疏特性,将多径水声信道冲激响应建为一个由各多径分量的时延和幅度组成的多径参数模型.该模型中输入信号向量产生的响应与多径时延参数呈非线性关系,与多径幅度参数呈线性关系.基于此特点,针对海洋水声信道的时变特性,分别采用进化算法和LMS自适应算法对模型中的多径时延和幅度参数进行混合寻优,从而解决时变信道条件下模型参数寻优困难的问题.仿真及海试信道跟踪实验结果表明:采用多径参数模型可降低模型寻优对象的维数,减少运算复杂性,提高估计效率;采用混合优化算法可减少多径参数模型的非线性寻优的复杂度,与进化算法相比,该算法具有更优越的时变信道跟踪性能.

As for the structural sparsity of impulse response of an underwater acoustic multipath channel,a multipath parameter model consisting of time delays and amplitudes of multipaths is proposed.The response of the multipath parameter model is nonlinear with the time delay parameters,but linear with the amplitude parameters.Based on this characteristic,as for the time-varying character of the underwater acoustic channel,a novel hybrid optimization method is developed to facilitate the optimization of the parameters of the time varying channel model;that is,the evolutionary algorithm and the least mean square algorithm are applied to address the model＇s time delay parameters and amplitude parameters,respectively.The results of simulation and sea-trial data demonstrate that the multipath parameter model can decrease the orders of the optimized parameters,reduce computational complexity and improve estimation efficiency.Besides,the hybrid optimization method can reduce the complexity of nonlinear optimization of the multipath parameter model,and compared with the classic evolutionary algorithm,it has a better tracking performance of the time-varying channel.