LMMSE based turbo equalization for nonlinear memory channel

摘要

Wideband satellite communication channel can be considered as nonlinear memory channel, which can be described by Volterra model. To counter both nonlinear and memory effects, linear minimum mean square error (LMMSE) based turbo equalizer is proposed in this paper. Although LMMSE is a kind of affine transform, the nonlinear memory interference is cancelled by taking the coefficients of nonlinear memory part into consideration. According to turbo method, the a priori probability of decision symbol is not taken into consideration in both linear part and nonlinear part. Meanwhile, full information rather than extrinsic information from the decoder is used to upgrade the performance. Simulation results demonstrate that the proposed LMMSE based equalizer has better performance than the original linear and traditional nonlinear equalizer.