Soft-output MIMO detector for MMSE receiver with channel estimation error

This paper considers the effect of channel estimation error (CEE) on the soft-output bit log-likelihood ratio (LLR) in a MIMO detection system, which the source sends the quadrature amplitude modulation (QAM) to the detector through Rayleigh fading channel. The new expression for the LLR is obtained for MMSE receiver using known pilot-symbol-aided channel estimation. Meanwhile, the optimal method for power allocation between training and data transmission, which is based on BER, is obtained under the total system transmitting power constraints. Numerical results show that the power allocation ratio sets to be 0.5 approximately is optimal. Moreover, the proposed LLR detector outperforms the conventional system without the consideration of channel estimation error. Meanwhile, the system performance which we have proposed is better than the existing research with the consideration of channel estimation.