MIMO Cube Decoder

An insight on the lattice decoder for flat-fading multiple antenna wireless communications systems is presented in this paper. In particular, we show that by formulating the decoding problem as a bounded-error subset selection, the resultant decoder finds the nearest lattice point to the received signal vector such that the search is bounded inside a hypercube centered at the received vector. The dimensions and orientation of the hypercube can be adjusted based on the diversity of the channel in order to improve its performance. The search for the nearest codeword to the received signal vector is solved by modeling the problem as an Integer Program (IP). Simulation shows that the proposed decoder is inferior to the Sphere Decoder (SD) by about 1-dB while its complexity is superior to the Sphere Decoder at very low signal to noise ratio.