Research On Simplification Of Multipath Channel Model And Sparse Channel Estimation

With the development of modern communication systems,the system bandwidth is increasing.The time resolution is improved,and the ability to distinguish different paths is enhanced.The number of resolvable paths in communication systems is particularly large,which makes the calculation simulation of channel models and the related applications extremely complex.Therefore,this thesis studies the simplification of the channel model under the premise of preserving the basic characteristics of the channel,expressing the dense multipath channel in a sparser and less-path form,or simplifying the existing channel model.In addition,with the development of communication technology,its application scenarios are also expanding,such as the high-speed railway,vehicle networking,drone communications,and so on.Many of these scenarios involve high-speed movement.The scenarios involving high-speed movement usually occur in open places,and usually multiple base stations serve users.At this time,each path can be regarded as a direct path,so that the number of the paths is small,showing sparsity in both time domain and Doppler domain.The user moves at high speed relative to the base station,and channel changes rapidly,resulting in a huge Doppler frequency offset,so that the conventional channel estimation algorithms are not applicable.This thesis studies the parameter estimation method of the high-speed channel with several large and opposite Doppler frequency offsets.The above two problems to be studied are essentially estimating channel parameters,including some or all of the parameters such as complex gain,path delay,and Doppler frequency offset.Therefore,these two problems can theoretically be solved in a similar way.In view of the problem of multi-path channel model simplification,this thesis first models the simplification of multi-path channel.Without considering the Doppler effect,using the delay-power spectrum to representation multipath channel on the time domain,the frequency correlation function is chosen as the feature of channel on the frequency domain.The error function is defined as the sum of the squared difference between the frequency correlation functions of original and simplified channel models.Thus,the simplification problem of the multipath channel model is transformed into the problem of seeking the minimum value of the error function.Next,we decompose the simplification problem into two sub-problems:1.How to calculate the power parameters when the delay parameters of each path of the simplified channel is known;and 2.How to determine the delay parameters so that the error function value is the minimum.As for the calculation of the power parameters in the simplification channel model,two methods are presented in this thesis:the least square(LS)algorithm and the Lagrange multiplier method.By analyzing the complexity and simulation results,it is concluded that although the complexity of the Lagrange multiplier method is slightly higher than that of the least square method,the estimation accuracy of the Lagrange multiplier method is better than that of the LS algorithm.So,the Lagrange multiplier method is used in the subsequent research to calculate the power parameters.As for the determination of delay parameters,this thesis analysis four methods:weighted merger method,exhaustive search algorithm,fixed-step search method and the improved particle swarm optimization algorithm.The first three methods are traditional and simple and are used as comparison schemes.The simulation results show that the weighted merger method has a low computational cost but the effect is poor.The exhaustive search algorithm has the best effect while it owns the highest complexity and is not applicable in the case of high dimensionality.The fixed step search method is optimized based on the rules summarized in the exhaustive search algorithm,which can reduce certain complexity,but it is currently only applicable to two-path simplified model.The improved particle swarm optimization algorithm has a reasonable calculation cost and the effect in the simulation application of the actual channel model Clustered Delay Line-C(CDL-C)and CDL-B channels is good.Therefore,it is a feasible multi-path channel simplification method to simplify channel power parameters and delay parameters by using Lagrange multiplication and improved particle swarm optimization algorithm,respectively.Then,in view of the problem of sparse channel estimation in high?speed scenes,two algorithms are proposed,the particle swarm optimization(PSO)algorithm of combining linear least square frequency offset estimation algorithm and the improved pure PSO algorithm.The former uses the particle swarm optimization algorithm to estimate the delay,and based on the obtained delay,the frequency offset is estimated using the linear least-multiplier frequency offset estimation algorithm.The latter uses particle swarm optimization algorithm to search for both delay and frequency offset.Both algorithms perform complex gain calculations after determining delay and frequency offset.Taking the linear least-multiplier frequency offset estimation algorithm with known correct delay as a comparison scheme,the simulation results show that the two algorithm effects are similar in estimation error,delay and frequency offset estimation accuracy to the comparison scheme.The pure PSO algorithm has better estimation error and frequency offset estimation accuracy than the comparison scheme when the different path has the same delay and the opposite direction of frequency offset.