Font Size: a A A

Research On Detection Technologies In The Receiver Of Mimo Communication Systems

Posted on:2014-02-23Degree:DoctorType:Dissertation
Country:ChinaCandidate:S L WangFull Text:PDF
GTID:1228330401463096Subject:Communication and Information System
Abstract/Summary:PDF Full Text Request
Because Multiple-Input Multiple-Output (MIMO) communication systems can achieve huge capacity or significantly enhance the transmission reliability, it has been widely used by the next generaion wireless communication systems. Detection in receiver is one of key technologies of MIMO communication system. The choices of detection technologies depond on the error performance and the computational complexity of the detection algorithms. Sphere detection (SD), which can significantly reduce the complexity of maximum likelihood detection (MLD) without losing its optimality, is a main method to achieve the tradeoff between the performance and the complexity. This paper researches SD to achieve a better tradeoff when MIMO is used as spatial multiplexing.1. A new initial radius choice algorithm based on the expected number of lattice points in hyper sphere is proposed. The complexity of SD depends heavily on the setting of initial radius which determines the volume of the hyper sphere. The proposed algorithm sets the initial radius based on the expected number of lattice points in hyper sphere so the proposed algorithm can get the probabilistic minimum.2. This paper proposes a new radius control algorithm which reduces the complexity of SD by early termination, i,e. when a candidate lattice point is obtained, we calculate the probability that the distance of correct lattice point is smaller. If the probability is small enough, SD will be early terminated. Though analysis and simulation, it is showed that the complexity reduction is significant while maintaining negligible performance degradation. 3. A tree pruning algorithm, named MMSE-PTSD, is proposed. MMSE-PTSD performs tree pruning based the solution to MMSE detection. Since the complexity of MMSE detection is much lower than MLD’s, the increase of complexity can be neglected if we preform MMSE detection before SD. Because MLD only concerns the lattice point which has the minimum distance, the path unlikely corresponds to the correct vector if the ratio of its path metric to MMSE solution’s path metric at the same layer is large. Then the path can be pruned. Though analysis and simulation, it is showed that the complexity reduction is significant while maintaining the limited performance degradation.4. Another tree pruning algorithm, PMP-SD, based MMSE-TPSD, is proposed. PMP-SD prunes a path according to the ratio of the path metric of current path to the minimum path metric at the same layer. Similarly, the complexity is significantly reduced with limited performance degradation. But MMSE detection is not necessary.5、Based on the reliability computation of MMSE decoder’s decision variable under BPSK modulation, two improved algorithms, Rule-R and Rule-M, which are applicable for any PAM constellation are proposed. Through appling the algorithms to channel matrix ordering of sphere detection and sub-optimal sphere detection, their complexities are significiantly reduced and performance of sub-optimal sphere detection can be much improved.6、Two channel matrix ordering algorithms base on Rule-R and Rule-M are proposed for FSD (Fixed-complexity sphere Detection). The proposed algorithm overcomes the sequential nature of FSD channel matrix ordering so it can make use of inherent parallelism of hardware platform and real-time performance can be significantly improved. Simulation shows that ordering algorithm based on Rule-R maintains the negligible performance degradation.
Keywords/Search Tags:multiple-input multiple-output (MIMO), spheredetection, radius control, tree pruning, fixed-complexity spheredetection
PDF Full Text Request
Related items