Research On Tree Searching Signal Detection Algorithm For MIMO System

Multiple-Input Multiple-Output (MIMO) makes a significant breakthrough in the field of wireless mobile communications in recent years. It sets up multiple antennas at the transmitting and receiving terminals respectively, and expands the one-dimensional smart antenna technology. Under the condition of not increasing the system bandwidth and total power of transmitting antennas, multi-antenna technology can make full use of space resources, resist the wireless channel fading effectively and improve the communication systems spectral utilization ratio and channel capacity. It is one of the key technologies in the new generation wireless mobile communication systems.Firstly, introduces the principle of multi-antenna wireless communication systems, the current research progress and its application in the field of wireless communication. It highlights to analyze the system model and system capacity. It analyses the traditional detection algorithms of MIMO system, and compares the performance of different algorithms according to the simulation. Then this paper analyzes the MIMO signal detection algorithm which has been widely studied in recent years based on tree searching, and studies the principle and searching process of the depth-first sphere decoding algorithm and breadth-first QRD-M algorithm. Finally, it summarizes this class of algorithms. The main research contents can be summarized as follows:(1) This paper studies the depth-first sphere decoding algorithm. It is found that the selection of the sphere radius is very important for the complexity and performance of the algorithm. Especially, the overlarge searching radius will lead to the complexity of the algorithm to reach maximum likelihood’s approximately when the signal-to-noise ratio (SNR) is very low. For this problem, it improves the traditional sphere radius according to using a compression factor and SNR to control the size of the sphere radius. It makes the complexity of the improved algorithm have obvious improvement in the low SNR. Experimental results demonstrate that the complexity of the improved algorithm reduces about 10% comparing with the traditional sphere decoding algorithm when the SNR is lower than 10 dB.(2) This paper also studies the breadth-first QRD-M algorithm. Because of the high redundant computation in the traditional QRD-M algorithm, the adaptive QRD-M algorithm is proposed. It uses a threshold value in each layer, and extends each layer partially. Experimental results demonstrate that the improved algorithm sacrifices little bit error rate and reduces the complexity greatly. So it can be better applied in the actual system.(3) The traditional QRD-M algorithm can discard the optimal paths easily in the process of detection, and do not consider the effect of noise, so it has a certain impact on its bit error rate. For these problems, QM-MMSE-SQRD algorithm is proposed. The new improved algorithm restrains the channel noise effectively, and uses the idea of sorted QR detection algorithm. Experimental results demonstrate that the new improved algorithm can enhance the performance of the traditional QRD-M algorithm effectively.