Research On Signal Detectionalgorithm For MIMO Systems

Multiple Input Multiple Output technique is regarded as one of the key techniques for next-generation mobile communication systems. It can enhance mobile communication systems capacity and performance, but its implementation complexity increase in the signal detection. It is a great challenge to reduce the computational complexity of signal detection in MIMO systems, so the signal detection method has high research value in MIMO techchnique. This thesis mainly researches on signal detection algorithms in MIMO systems. We study sphere decoding algorithm and give an improved scheme to acheieve a tradeoff between performance and complexity.Firstly, the wireless channel model and the fading characteristics of it are introduced, then the model for MIMO systems. And it compares several traditional signal detection algorithms for MIMO systems. Then Sphere decoding algorithm has optimal error performance and medium computational complexity, compared to maximum likeihood detection and other traditional signal detection. However, the exist sphere decoding algorithm can not solve the systems where there are more transmit antennas than the receive antennas. The exist generalized sphere decoing algorithm can overcome this shortcoming, but the computational complexity is very high.Secondly,λ- GSD algorithm can not only approach the maximum likelihood decoding in error performance and provide significant reduction in computational complexity, but also can be directly applicable to any signaling constellation. Then this paper combineλ-GSD algorithm with ZF-OSIC detection algorithm to offer an improved algorithm, which detects high-reliability symbols with ZF-OSIC before the other symbols withλ- GSD. The simulations indicate that the improved algorithm has optimal performance and medium complexity.Finally, sphere decoding with compressive sampling is introduced. It can capitalize on both the sparsity and the finite-alphabet features of a sparse finite-alphabet signal. For its shortage of the initial radiu selection, this paper proposes a new initial radiu selection method based on orthogonal matching pursuit algorithm. The simulations show that the calculation complexity of the improved method can outperform sphere decoding with compressive sampling.