一种基于遗传算法的新时延估计方法

摘要

提出了一种新的时延估计方法，通过采用FIR滤波器模型并结合遗传算法解决了代价函数复杂计算量下的高效全局优化。算法利用最小二乘法准则，推导得到优化目标函数，并将时延，滤波器系数列入到参数估计模型中，继而将目标函数作为适应度函数，将时延，滤波器系数作为决策变量，应用遗传算法进行全局优化。通过仿真实验表明，本文不仅在滤波器长度比较短的情况下获得最优的时延估计，而且大大地减少了计算量，提高了稳健性。%A novel time delay estimation approach is proposed. It solves efficient global optimization via FIR filter model with genetic algorithm under the complex calculation of cost function. In the described method, objective function is derived by use of least-square-based rule; time delay parameter and filter coefficients are included into the parameter estimation model; then objective function is converted into fitness function and time delay as well as filter coefficients are considered as decision variable; global optimization is operated via genetic algorithm. Simulation experiments show that the proposed method can acquire optimal time delay estimation in the case of short filter length, but also reduce the computation and improve the robustness.