Research On Multi User MIMO Detection Technology Of LTE Relay System

Relay technology, one of the key technologies of LTE system, is to increase the relaynode (RN) is based on the existing site, the source node to destination node data can passthrough a relay node, so as to improve the transmission rate, expand the area coverage andsystem capacity. In order to more effectively improve the data transmission rate and capacityof the system, relay technology to integrate with a variety of techniques, especially multiinput multi output (MIMO) technology. Due to the restriction of the terminal equipment bytechnology and hardware, not possible to install multiple antennas, hopes multiple usercooperative communication, work together to form a virtual MIMO system, MU-MIMOsystem. In recent years, multi user detection algorithm is the focus of the industry, this paperresearches on it in LTE relay system.First, analyzes the evolution process of LTE relay system architecture. Analysis ofdifferent development stages of LTE relay architecture, points out the existing problems ofLTE relay system, finally forecast the development trend of relay.Then, simulation analysis the performance of the traditional detection algorithm inMU-MIMO system. Compared with zero forcing (ZF) algorithm, minimum mean squareerror (MMSE) algorithm and the ordered successive interference (OSIC) algorithm,,ML-OSIC algorithm has advantages in terms of performance. But compared with the optimaldetection algorithm, ML-OSIC algorithm are still gaps.Finally, the application can improve the lattice channel Orthogonality (LLL) technologyin the ML-OSIC detection algorithm, ML-LLL-OSIC algorithm is proposed, and simulationanalysis by MATLAB. The simulation results show that: in a high signal to noise ratio,compared with the performance of ML-OSIC algorithm, ML-LLL-OSIC algorithmsignificantly improves the detection performance; but the low signal-to-noise ratio detectionperformance of the ML-LLL-OSIC algorithm under not ideal. For the low SNR circumstance,this paper further to improve the ML-LLL-OSIC algorithm, the improved algorithm combines the advantages of ML-OSIC and ML-LLL-OSIC of two algorithms, for both highand low SNR application.