A low-complexity transceiver scheme based on minimum Euclidean distance criterion

A low-complexity maximum likelihood (ML) transceiver scheme is investigated for multiple-input multiple-output (MIMO) systems without any constraint on both the number of data streams to be transmitted and the modulation order for quadrature amplitude modulation (QAM). The max-d_min precoding problem is first reconsidered in the presence of rank constraint on the so-called precoder generating matrix. Relying on the channel combination factor, an efficient uniform framework is then presented to construct suboptimal high-dimensional precoder generating matrix via lower-dimensional ones. To further alleviate the complexity introduced by highorder QAM input, 4-QAM is treated as the unique constellation at the expense of increased number of data streams. In addition, benefiting from the special structure of the proposed precoder, a new complexity-reduced ML decoder is proposed with particular permutation matrix. Numerical results indicate the flexibility and superior performance of the proposed transceiver over the optimal MMSE transceiver in terms of bit error rate (BER).