基于GSA-BP神经网络的OFDM系统信道估计算法

针对正交频分复用(orthogonal frequency division multiplexing， OFDM)系统中存在的不可忽视的非线性噪声问题，为了能够更好了解信道特性，需要利用信道估计获得信道状态信息，提出一种基于黄金正弦优化BP(golden sine algorithm，GSA-BP)神经网络的OFDM系统信道估计算法，克服了传统 BP神经网络算法容易陷入局部极值的问题，提升了信道估计算法的估计精度.首先通过LS信道估计算法获得信道的初始估计，再将其通过GSA-BP神经网络算法得到信道的精确估计.仿真结果表明，在相同的信道环境下，提出的算法比LS算法具有更好的性能，与MMSE算法性能接近，但不需要信道先验统计特性，易于实现.

Channel Estimation Algorithm of OFDM System Based on GSA-BP Neural Network

The nonlinear noise problem in orthogonal frequency division multiplexing(OFDM)system cannot be ignored. In order to understand the channel characteristics better， channel state information is needed by channel estimation. A channel estimation algorithm for OFDM system was proposed based on golden sine algorithm optimized BP neural network(GSA-BP). The problem was overcome that the traditional BP neural network algorithm is easy to fall into local extremum， and the estimation accuracy of channel estimation algorithm was improved. Firstly， the initial estimation of the channel was obtained through the LS channel estimation algorithm. Then， the accurate estimation of channel was obtained by using GSA-BP neural network. Simulation results show that the proposed algorithm has better performance than LS algorithm， and is close to MMSE algorithm， but it does not need channel prior statistics and is easy to implement.