A Study On Channel Estimation Algorithm In High-speed Mobile Environment

Orthogonal Frequency Division Multiplexing (OFDM) is a parallel transmission scheme. Because of its high spectrum utilization ratio and better resistance to multipath interference, OFDM has been widely used in the field of wireless communications. High-speed mobile environment will lead to large Doppler frequency shift, which will destroy the orthogonality of OFDM system and cause severe Inter-Carrier Interference (ICI). ICI will seriously affect the reliability of communication system, which makes the channel estimation no longer accurate. Therefore, channel estimation is considered as a critical problem in high mobility environment in OFDM system.In this dissertation, we analyse the propagation characteristics of wireless channel, especially focus on the statistical channel model and high mobility wireless channel characteristics. By studying the principles of OFDM, the communication system model is proposed. Then, the traditional pilot-aided channel estimation algorithm and estimate criterion are analyzed. The channel due to the high mobility is assumed to be time-varying, which means the channel can not be simply described as a linear time-varying channel model. At this point, the traditional frequency domain aided channel estimation will lead to serious error floor. In order to reduce the estimation error and achieve better performance, Basis Expansion Model (BEM) channel estimation method would be a good choice. Instead of the traditional method of channel estimation, only some finite model parameters need to be estimated by using BEM pilot-assisted method. However, the estimation of parameters is seriously affected by ICI from non-pilot symbols. Based on the analysis of ICI power distribution, it's easily to be seen that subcarrier interference mainly comes from the adjacent subcarriers. To reduce the ICI, we put forward a new way of pilot called group-pilot. In the dissertation, the power distribution of ICI is derived firstly. And then, group-pilot is proposed to estimate the BEM parameters based on frequency pilot-aided method. Finally, the simulations show better performance than the comb-type BEM methods in fast fading channel.Finally, this dissertation analyzes the traditional block-type pilot-aided channel estimation methods. For fast fading environment, mentioned above in the traditional channel estimation algorithm is no longer applicable. For this, an improved method based on short block-type pilot-aided time-varying channel estimation algorithm is proposed. The new algorithm aims to solve the problems that pilot symbols are under severe ICI and frequency domain interpolation is of great error in time-varying channel. And simulation results indicate that, under the high speed mobile environment, the improved algorithm show better performance in the time-varying channel.