Design Of IP Module Based On Wireless Communication In MIMO Signal Detection

With the development of semiconductor technology,microelectronics technology and the computer technology,wireless communication technology has achieved the rapid development and obtained wide application and it constantly improves the people's way of life,which makes people's life into a new stage of development.This paper first presents the development of the communication industry,and then analyzes the performance and complexity of several traditional sign detection in Multiple Input Multiple Output(MIMO)system.By studying the traditional Sphere Decoding,unfortunately the traditional Sphere Decoding offers an intrinsically variable throughput,which tends to decrease significantly at low signal-to-noise ratio,another disadvantage of the traditional Sphere Decoding is the sequential tree search order,which makes it difficult to employ parallel architectures to improve the throughput.Then we proposed an improved sphere decoding algorithm on the basis of the traditional Sphere Decoding.We first get the smaller sphere by a suboptimal detection algorithm,then we propose a novel four-nodes-per-cycle parallel architecture that exploits the breadth-first visiting order to improve throughput and shorten the critical path.And finally we simulate the improved algorithm simulation to get the BER performance with MATLAB.At last,we implemented the main functional module of improved sphere decoding algorithm in FPGA and ASIC technology,including completing the physical backend design with EDA tools and finishing Package IP successfully.Compared with the traditional signal detection algorithm,the improved SD algorithm can get a max clock frequency of 345 MHz,while saving 25% in area cost,it also get a good performance in signal detection.The main works of this paper are as follows:1)We built single-user and multi-user MIMO-OFDM wireless communication baseband signal processing simulation link based on the existing conditions,and established future-oriented wireless communications baseband signal processing experimental platform successfully,we also provide the verification platform for the baseband signal detection algorithm simulation.2)An improved SD detection algorithm is proposed based on the traditional SD detection algorithm,firstly we derived a smaller sphere radius by pre-process through the sub-optimal detection algorithm,And then we use a four-point cycle of a parallel structure,This structure is based on a width-first search strategy that can greatly improve data throughput while reducing largely average complexity of the algorithm..3)The improved SD detection algorithm is implemented in FPGA and ASIC respectively to keep the distance with traditional SD algorithm,we measure different working frequency under different crafts.At the same time,the improved SD detection algorithm is implemented by back-end EDA tool,the results show: compared with the traditional SD detection algorithm,the proposed solution can greatly reduce the area of hardware resources,when getting the closest performance of the ML algorithm detection,we also achieve a higher data throughput.At last,it provides an important guidance and reference for the further application of the improved SD detection algorithm to the design of a special processor for communication.