Multi user detection of overloaded mimo systems using network coding

Abstract

efficient by deploying multiple antennas as both side of a communication link thereby called Multiple Input Multiple Output (MIMO) arrangements. But once these multiple arrangements are integrated with the system, co channel interference increases and it becomes almost challenging to provide comparable number of receive antennas and transmit antennas, hence creating the popular over-loaded (rank deficient) wireless systems. In this situation, Maximum Likelihood (ML) works well but its complexity raises exponentially with the increase in number of transmit antennas whereas other linear detection algorithms do not perform well. In this thesis, a new signal processing technique for multi user overloaded MIMO systems using a combination of Network Coding (NC) with Hybrid Automatic Repeat reQuest (HARQ) protocols have been examined. The re-transmissions are stacked using HARQ to create virtual receiver antennas which transmutes an overloaded system into a critically loaded system (i.e. a system having equal number of transmit and receive antennas) while NC protocol has been used to improve the throughput of the channel as well as efficient bandwidth utilization. In the start, an overloaded multi-user system model is considered where it is demonstrated that the Minimum Mean Square Error (MMSE) Detector performs poorly as compared to Maximum Likelihood (ML) Detection for such systems. To overcome this drawback, a novel transmission approach is proposed that works well under overload condition while utilizing the channel bandwidth efficiently and providing more capacity to the same. This approach allows us to use linear Multi-user Detection (MUD) algorithms without mounting additional antennas chains physically in the system’s hardware. Monte Carlo Simulations prove that the proposed approach results in significant gains in terms of Error Rates and throughput of the channel while keeping the dropped packet rate quite low.