Symbol Detection Techniques in a Spatial Multiplexing System

Abstract

Significant performance gains are achievable in wireless communication systems using a Multi-Input Multi-Output (MIMO) communications system employing multiple antennas. This architecture is suitable for higher data rate multimedia communications. One of the challenges in building a MIMO system is the tremendous processing power required at the receiver side. MIMO Symbol detection involves detecting symbol from a complex signal at the receiver. The existing MIMO detection techniques can be broadly divided into linear, non-linear and exact detection methods. Linear methods like Zero-Forcing offer low complexity with degraded Bit Error Rate (BER) performance as compared to non-linear methods like VBLAST. Non-linear detectors are computationaly not very expansive with acceptable performance. Exact solutions like Sphere Decoder provide optimal performance however it suffers from exponentional complexity under certain conditions. The focus in the early part of this thesis is on non-linear approximate MIMO detectors and an effort has been made to develop a low complexity near-optimal MIMO detector. Computational Swarm Intelligence based Meta-heuristics are applied for Symbol detection in a MIMO system. This approach is particularly attractive as Swarm Intelligence (SI) is well suited for physically realizable, real-time applications, where low complexity and fast convergence is of absolute importance. Application of Particle Swarm Optimization (PSO) and Ant Colony Optimization (ACO) algorithms is studied. While an optimal Maximum Likelihood (ML) detection using an exhaustive search method is prohibitively complex, we show that the Swarm Intelligence optimized MIMO detection algorithms gives near-optimal Bit Error Rate (BER) performance in fewer iterations, thereby reducing the ML computational complexity significantly. In this thesis novel non-conventional MIMO detection approaches based on Swarm-Intelligence techniques have been presented. An effective and practical way to enhance the capacity of MIMO wireless channels is to employ space-time (ST) coding. Space-time block coding (STBC) is a transmit diversity technique in which the data stream to be transmitted is encoded in blocks, which are distributed among multiple antennas and across time. Alamouti’s simple STBC scheme for wireless communication systems uses two transmit antennas and linear maximum-likelihood (ML) decoder. This system was generalized by Tarokh to an arbitrary number of transmit antennas by applying the theory of orthogonal designs. In the later part of this thesis a simple multi-step constellation reduction technique based decoding algorithm that further simplifies the linear ML detection in Orthogonal Space-Time Block Coded systems is proposed. This approach reduces the computational complexity of these schemes while presenting the ML performance. In addition, Spatial Multiplexing systems using Orthogonal Walsh codes are also studied. This approach has a potential to reduce the search space to allow efficient symbols detection in Spatial Multiplexing systems.