Research On Doppler Frequency Shift And Network Time Delay Estimation In High-speed Scene

In high-speed scenarios,fast moving speed can cause a lot of problems,for example,the increasement of Doppler frequency shift and transmission delay.When the user is in high speed motion,because of the increase of Doppler frequency shift,the orthogonality between subcarriers in OFDM system will be destroyed.Thus the accuracy of channel estimation decrease,leadings to the degradation of the system performance.It needs a certain adaptive adjustments to adapt to the changes in the communication environment based on the current Doppler frequency shift conditions,hence,the estimation accuracy of the network delay,which is highly affected by the transmission distance,the network load and the jitter of the time delay,is very important to the performance of the system.Users in high-speed movement will experience the change of transmission delay,however,most of the network time delay estimation studies in nowadays are based on the historical information of time delay,without considering the influence of network load and the change of moving speed.These time delay mutations will result in certain estimation errors.In order to solve the problem of Doppler frequency shift estimation and network time delay estimation,this thesis performs a simulation study on these problems based on Kalman algorithm.In this paper,a link level simulation platform is built,on which the communication process of physical layer is realized with,four kinds of pre coding methods,including pre-coding based on ZF Criterion,pre-coding based on MMSE Criterion,pre-coding based on SVD decomposition and pre-coding algorithm based on GMD decomposition.The precoding scheme based on the ZF criterion can reduce the interference between the channels very well,and the precoding algorithm based on the MMSE criterion introduces the system noise into the algorithm and reduces the influence of the system noise.Theoretically speaking,the precoding algorithm based on MMSE criterion should be more excellent.The GMD-based precoding algorithm allocates more power to sub-channels with poor performance,so that the performance of each channel is quivalent,so it is better than the performance of precoding based on SVD decomposition.Compared with the advantages and disadvantages of several methods,the accuracy of the LTE link level simulation platform is validated,which provides a good simulation environment for the follow-up research.For the estimation of Doppler frequency shift,this paper firstly simulates the Doppler frequency shift estimation algorithm based on phase difference method,then proposes an improved algorithm based on Kalman filter,which overcomes the phase difference method in high speed motion with less estimation errors.The adaptive pilot interval based on Doppler frequency shift is proposed,which can adapt to the change of channels.For the estimation of network delay,this paper firstly estimates the network delay based on extended Kalman filtering algorithm and square root Kalman filtering algorithm,then discusses the influence of moving speed,network load and delay jitter on estimation.The Kalman algorithm based on multivariate linear regression is proposed to eliminate the influence of the above factors on the estimation error.