Simulation And FPGA Implementation Of MIMO Detection

The demands on data transmission rate and communication performance increase with the rapid developments in wireless communication technology. Due to the limited resource of wireless spectrum, people pay more attention on Multiple Input Multiple Output (MIMO) technology which is a major breakthrough in mobile communications technology. In wireless communication system, as long as multiple antennas are used at the transmitter and receiver, a wireless MIMO system can be obtained. Since the MIMO technology can improve the system capacity and spectrum efficiency significantly without increasing the bandwidth and convert the traditionally considered harmful factors that are caused by multipath fading into positive ones, it has become a key technology in next generation wireless communication systems.This thesis firstly introduces two categories of MIMO technology, i.e. MIMO diversity system and multiplexing system. On the basis of in-depth analysis of MIMO technology, the thesis focuses on the MIMO detection algorithms with simulation comparisons among various algorithms. Based on analysis of the simulation results and the complexity of hardware implementation,the thesis ultimately determines to use the MMSE(Minimum Mean Square Error) algorithm as the hardware implementation method. As an important and complex module in MIMO detection algorithm, the matrix inversion is analyzed on the basis of some commonly used matrix inversion algorithms so as to find an efficient and low complexity algorithm.Secondely, the thesis focuses on the hardware implementation of MIMO detection algorithms. The FPGA development platform is introduced at the beginning with the chip selection. According to the theoretical data, the data fix point analysis in performed and the FPGA module of MIMO detection is implemented,which includes matrix computation module, complex matrix inversion module, minimum Euclidean distance and constellation point positioning module and so on. We use the language Verilog to implement the modules and XILINX ISE as the functional simulation tool to verify the correctness of the implementation. Especially, a high-effective algorism of the reciprocal module which can greatly reduces the complexity of the module is proposed and used to implement the complex matrix inversion module. Both the test results and performance analysis are provided.The thesis concludes with the summary of the main work as well as the suggestions of future work.