Lattice reduction aided pre-processor for large scale MIMO detection

Abstract

Linear detectors such as Minimum Mean Square Error (MMSE), Zero Forcing are applied for large scale MIMO detection because of their low computational cost, but has the least performance. In this paper, improvement in the performance of lattice Reduction aided- linear detectors for large scale MIMO is analyzed. Lattice reduction is performed using Seysen's Algorithm (SA). SA finds the reduced basis with minimum number of iterations compared to the commonly used Lenstra, Lenstra, and Lovász (LLL) algorithm. Also, SA based reduction of channel matrix, outperforms the existing LLL algorithm based reduction for sub optimal linear detectors in terms of bit error rate (BER) and computational complexity. The simplified VLSI architecture of SA is implemented in hardware using 45 nm technology