Enhanced lattice-reduction-based precoder with list quantizer in broadcast channel

Abstract

A precoding scheme for multiuser broadcast communications is described. In the multiple-antenna broadcast channel, each user is not allowable to cooperate other users, so the efficient and well-designed precoding scheme to reduce other user’s interferences is needed. As a simple precoding, linear precoding using linear pre-equalization filter at transmitter was proposed [11]. Furthermore, to achieve the capacity of the multiple-antenna broadcast channel, studies have shown that combination of dirty paper coding [3] and linear precoding [13], using spheresearching algorithm. Recently, the use of lattice basis reduction(LR) has been shown to be a very effective method for precoding in multiuser multiple antenna systems [14]. Lattice basis reduction helps us to reduce the average transmitted energy by modifying the region which includes the constellation points. Even though the LR precoding method achieves a full diversity order, the performance loss caused by quantization error still remains. In this thesis, we propose a simple precoding technique which adapts lattice basis reduction and reduces quantization error, analyzing the patterns of the error. The proposed scheme establishes a lattice list to provide candidates for more reliable lattices due to the quantization error with low computational complexity. Simulation results show that the proposed scheme performs well, and as compared to the more complex methods (such as the perturbation vector using sphere-encoding) has a negligible loss. Furthermore, the complexity of these schemes are compared and proposed scheme has still low complexity with performance gain.