Research And Implementation Of Physical Layer Technology For High Speed Communication System

The MIMO-OFDM system combines MIMO technology and OFDM technology.It has the characteristics of high robustness,high data transmission rate,and high spectrum utilization rate.It has achieved a significant breakthrough in the field of traditional communications.However,due to the complexity of the wireless communication transmission environment,the phase and amplitude of the received signal will be severely faded,and the transmitted data must be recovered through channel estimation technology.Therefore,considering both accuracy and computational complexity,it is very meaningful to develop a MIMO-OFDM channel estimation scheme suitable for hardware implementation.The main work of this article is as follows:(1)Two key points of MIMO-OFDM channel estimation technology are studied,the pilot-based channel estimation technology and the choice of appropriate training sequences.In terms of channel estimation technology,the commonly used channel estimation algorithms are studied in detail: LS,MMSE,LMMSE,and SVD-LMMSE and interpolation algorithms: the basic principles and hardware implementation complexity of linear interpolation,Gaussian interpolation,and cubic spline interpolation.The channel estimation algorithm,pilot method,and interpolation method are simulated under different channels,which verifies the rationality of system selection of block pilot and LS channel estimation under the two factors of complexity and performance.In terms of the selection of training sequences,different channel estimation algorithms are designed and simulated for the preamble part and data part according to the standard 802.11 ac physical layer frame structure and transmitter structure.For the estimation of the preamble part,two L-LTF training sequences are averaged in exchange for 1dB performance improvement;for the data part,the VHT-LTF training sequence is used for channel estimation based on the LS criterion,and then the phase offset correction is performed for each OFDM symbol by by inserting certain pilot subcarriers.Simulation proves that the algorithm can achieve excellent performance under different MCSs,different frame lengths,different channels and different channel frequency offsets.Then the design of the training sequence in 802.11 ac and the channel estimation algorithm are applied to the MIMO-OFDM system of this design.A LS channel estimation algorithm based on training symbols and an algorithm for channel tracking correction by inserting pilot symbols in the data part is proposed.Simulations prove that the algorithm will not deteriorate with the increase of phase deviation.(2)In the 40 MHz bandwidth,40 MHz sampling rate,and 125 MHz clock frequency,an 8-transmit and 8-receive MIMO-OFDM system with a load rate of 1Gbps is implemented on the FPGA.The FPGA implementation of the framing module and the channel estimation and detection module is analyzed and verified by RTL simulation,and the error of the channel estimation module implemented on the FPGA is analyzed and compared with MATLAB.The function correctness and throughput of the module is determined under the requirements of high-speed transmission.During the test on the board,the method of combined hardware and software testing is used.The data interaction between MATLAB and FPGA is realized through the LWIP protocol stack.The performance of the system in MIMO channels is quantified,and its performance requirements and goals are verified to be consistent.