Research On Detection Algorithm Based On Lattice Reduction In LTE-A Downlink

In order to meet user requirements, MIMO-OFDM technology is adopted in LTE-A downlink to improve the spectrum efficiency, system capacity and data transmission rate without increasing system bandwidth.Due to uncertainty of the wireless channel, the transmission data is influenced by multipath effect and Doppler shift of wireless channel.Furthermore, it is also affected by the interference between antennas seriously. Therefore, it is need to design high-performance signal detector at the receiver to ensure the reliability of data transmission. The lattice reduction aided detection algorithm is widely used in LTE-A for unlimited approximate maximum likelihood detection algorithm performance.This paper studies the lattice reduction aided detection algorithm in LTE-A downlink from the following several aspects.This paper first studies conventional signal detection algorithms in LTE-A downlink,which provides theoretical basis for research in subsequent chapters. To overcome error diffusion and low diversity of traditional QR decomposition algorithm, a cyclic iteration SQRD detection algorithm based on PIC preprocessing is proposed. The improved method combines the thought of sorting and loop iteration,using PIC to update decision value in every detection process to improve the detection precision in target layer.Then, this paper studies real lattice reduction aided detection algorithm, for the problem that signal detection performance of conventional algorithms severely decrease and cannot get full diversity when LTE-A system is in high speed environment. This paper proposes a lattice reduction aided detection algorithm based on noise prediction correction. Combining thoughts of noise prediction and ML, noise for undetected signal component is predicted using detected signal component. The influence of noise can be eliminated before signal detection.Finally, for the problem that complexity of real lattice reduction aided detection algorithm is too high when the LTE-A system is in low speed environment, this paper studies complex field lattice reduction aided detection algorithm. A lattice reduction aided parallel detection based on quantization error correction is proposed to overcome error diffusion which caused in quantization process. The branches is formed when the layer which is most severely affected by noises experiences all possible signal values. Then proceeding lattice reduction aided detection for residual signal in every branches and modify the quantitative values in quantization process to get estimate values. Finally using ML to choose the optimal decision value. Simulation results show that the improved algorithm overcomes error diffusion and detection performance reaches the approximate optimal.