Research On Reconstruction Algorithm Of Millimeter Wave Channel Estimation Based On Compressed Sensing

As a popular research direction in the current communication field,5G communication systems have different channel characteristics compared with those of previous generation communication systems.The popular research frequency band for 5G communication channels in the world is the millimeter wave frequency band.This dissertation focuses on channel estimation based on compressed sensing in the millimeter wave frequency band.The focus is on channel reconstruction algorithms in channel estimation based on compressed sensing.On the basis of existing algorithms,the accuracy and iteration speed of channel estimation are further improved.This dissertation first briefly introduces the wireless channel model,the OFDM system and the LS channel estimation.After completing the above work,the basic theory and specific algorithm flow of compressed sensing technology are briefly explained.At the same time,how to use compressed sensing technology to realize channel estimation in OFDM system is introduced.After clarifying the pilot structure suitable for channel estimation based on compressed sensing,this dissertation introduces a classic signal reconstruction algorithm in compressed sensing—OMP algorithm.The simulation results highlight the technical advantages of channel estimation based on compressed sensing over LS channel estimation algorithms.Secondly,after giving the specific modeling process of the millimeter wave multipath channel model used in the simulation experiment,the OMPBR algorithm is introduced.It can use binary search to further refine the estimated time delay of the OMP algorithm,so as to achieve higher precision channel estimation.Aiming at the slow iteration speed of the binary search algorithm in the OMPBR algorithm,this dissertation proposes a fast time-delay positioning algorithm based on the correlation value according to the correlation value characteristics of the pulse signal.The algorithm improves the positioning speed of time delay.It is proved by simulation that the OMPBR algorithm improved by the fast time delay positioning algorithm based on the correlation value does effectively reduce the number of conversions used in the positioning delay,and has no obvious impact on the accuracy of the estimation.Finally,in view of the current situation that most channel estimation algorithms improve the accuracy of channel estimation by improving the accuracy of time delay positioning,this dissertation proposes a dual-path reconstruction algorithm.The idea of the dual path reconstruction algorithm is to use the artificially designed second delay point and the first delay point estimated by the algorithm to work together to estimate the channel.Ideally,the second delay point can compensate for the deviation of the first delay point to a certain extent.Analysis and simulation experiments prove that the dual-path reconstruction algorithm effectively improves the accuracy of channel estimation compared with the OMPBR algorithm.Compared with the OMPBR algorithm,the improved dual-path reconstruction algorithm based on the correlation value delay location algorithm not only improves the estimation accuracy,but also reduces the number of iterations.