Study on Adaptive Channel Estimation in OFDM Systems Based on DSP/FPGA

摘要

In order to solve channel estimation problem in OFDM (orthogonal frequency division multiplexing) systems, a decision–directed method is proposed for channel estimation and equalization based on OFDM. The proposed method uses a received subcarrier-symbol to represent the competitive neural network, uses a nerve cell of the neural network to estimate adaptively the variation weight value, and then estimates the channel according to the channel weight value. Finally the channel equalization is reached by removing the constellation symmetry to compensate the nonlinearity effect. This method does not require the prearrangement knowledge about the wireless communication channel. Simulations and actual test show that the proposed method achieves very good channel estimation and performance equalization.