基于遗传算法的OFDM系统导频设计

Genetic algorithm based pilot design for OFDM systems

针对OFDM系统中虚载波破坏等间隔等功率导频最优性的情况,采用遗传算法设计了OFDM系统的导频信号.在给定导频个数的情况下,首先推导了最小二乘准则下信道估计均方误差(MSE)的一个上界,然后将其作为遗传算法的适应度函数,对导频的位置和功率进行优化.该适应度函数可有效逼近信道估计的MSE,且无需矩阵求逆运算.仿真结果表明:给定导频功率,导频位置的选取对MSE的影响较大;给定导频位置,导频功率对MSE的影响则相对较小;采用所提出算法设计的导频信号优于等间隔等功率的导频信号.

A genetic algorithm based pilot design for OFDM （orthogonal frequency division multiplexing ） system is proposed because of the fact that the existence of virtual carriers destroyed the optimality of the equal spaced and equal powered pilot tones. With given number of the pilot tones, an upper bound of the mean square error （MSE） of the least squares channel estimation is derived, and then the upper bound is used as the fitness function of the genetic algorithm in order to optimize the locations and powers of the pilots. The proposed fitness function can well approach the MSE and avoid the computation of matrix inversion. Simulation results show that the locations of the pilots have high impact on the MSE, while the powers of the pilots have low impact on the MSE. Furthermore, the pilot signal designed by the proposed approach is superior to the conventional equal spaced and equal powered pilot signal.