Research On The Sphere Decoding Method In Mimo Systems

Multiple-input multiple-output (MIMO) technique, which deploys multiple transmit antennas and multiple receive antennas, is a huge breakthrough in the field of wireless communications. It can exponentially increase the channel capacity and spectrum efficiency without the increasing of the bandwidth, and becomes one of the key techniques for the third generation and future mobile communication systems to achieve high data rate, high transmission quality and high system capacity.The quality of a communication system largely depends on the accuracy of the detection signals, while traditional signal detection algorithm can’t meet the requirements either because the performance deteriorates too fast or because the complexity of the algorithm is too high, the emerging semidefinite relaxation detection (SDR) algorithm made a good compromise between performance and complexity, but there always exists a gap between the optimal performance of the Maximum Likelihood Detection and SDR’s, and the gap will become larger and larger with the increasing of the antenna dimensions and the modulation order. Sphere decoding can greatly reduce the complexity of the Maximum Likelihood Detection and maintain the optimal performance in the same time.Nowadays, sphere decoding’s problem is that the complexity of the algorithm is too high when the SNR is too low or the antenna dimension is too large. When the SNR is low enough the difference between the optimal performance and the suboptimal performance is little, in such a case, the signifi-cance of sphere decoding is not that meaningful. Large MIMO systems have attracted much attention because of its great capacities, so, the focus of our thesis is to reduce the complexity of sphere decoding in large MIMO systems.This thesis studied the performance of sphere decoding with the increasing of the antenna dimensions in different SNR by union bound and found the separate point. The choice of the initial radius is one of the important factors that affect sphere decoding’s complexity. This thesis developed two improved schemes in large MIMO systems based on the optimal initial radius option. The first scheme which combined the radius sequence and binary search is an improvement of Minimum Mean Square Error (MMSE+SD) radius search algorithm. The second scheme is also an improved version of SDR, in this scheme, we just replace MMSE with SDR to narrow the scope of the search and reduce the complexity of sphere decoding. After the deep discussion of the first scheme, this thesis combined the two schemes to further reduce the complexity of sphere decoding.