Adaptive joint nonlinear transmit-receive processing for multi-cell MIMO networks

Abstract

This paper considers signal processing algorithms for the downlink coordinated multi-point multiple-input multiple-output (MIMO) systems with multiple-antenna users. A novel adaptive joint nonlinear transmit-receive processing algorithm is proposed based on zero-forcing (ZF) criterion. In this algorithm, a block successive ZF precoding is first applied at the multiple base stations to pre-cancel partial multiuser interference. Then nonlinear Tomlinson-Harashima precoding is applied to further reduce interference between the data streams of other users and the other data streams of the same user, whereas linear equalization and modulo operation are applied at each user. We first show that the proposed joint nonlinear transmit-receive processing algorithm effectively decomposes the multiuser MIMO channel into parallel independent single-user MIMO channels. Different from the method in [1], our proposed algorithm allows the number of sub-channels of each user to be arbitrary number no more than the rank of that user's equivalent channel, and then for the given set of the numbers of sub-channels and the given power allocation of all users, closed-form expressions of the transmit and receive processing matrices are derived to optimize the output signal to interference plus noise ratio of each user. Subsequently, an adaptive method is proposed to adapt the number of sub-channels according to the fading of the users' channel matrices (termed as adaptive algorithm). Simulation results show that the proposed algorithm achieves much better throughput performance than the processing methods in the literature, such as the non-adaptive nonlinear preprocessing algorithm in [1] and the known block diagonalization techniques.