Iterative Channel Estimation Based On Correlation Algorithm In Wireless Multipath Channel

In terrestrial wireless communication systems,the transmitted signal may be seriously distorted due to various obstacles,scatterers,and other sources on the propagation path.Under the influence of multipath effects,Gaussian white noise,co-channel interference and other factors,the received signal may be degraded,which may affect the accuracy of data transmission.Among the interference,the inter-symbol interference caused by the multipath effect is a key factor which leads to the signal degradation.Based on channel estimation,equalization technique is an effective approach to eliminate the inter-symbol interference in practice.Thus,the accuracy of channel estimation is directly related with the system performance.Under severe interference with low SNR,some common data-aided estimation algorithm like least square(LS)algorithm,least mean square(LMS)algorithm is not effective since the signal is totally submerged by noise.In this case,a fast correlated channel estimation algorithm with high anti-noise ability and processing speed is proposed.However,due to the effect of interference factor,such as noise,some estimation errors are introduced in the process of fast correlated algorithm,and these errors will affect the accuracy of estimates.In this thesis,the fading and related parameter characteristics in wireless channel are studied in the first section.Simulations and analysis are also implemented on the different fading models.Then,some types of classic models under single-input single-output(SISO)channel are studied in the next part of this thesis,including indoor channel model and outdoor channel model.In these models,exponential model is a useful model to deal with the multipath delay,thus this thesis chooses this model to research on channel estimation.On the basis of these works,two kinds of iterative channel estimation algorithms are proposed to solve the problems in the existing estimation algorithms.Based on the fast correlated algorithm,both of the algorithms can reduce estimation errors of the original algorithm and improve the accuracy of the estimated results.What's more,the corresponding mathematical conclusions,for example,the relationship between the estimates with the number of iterations,complexity analysis and the Cramer-Rao bound of mean square error are deduced to support the proposed algorithms.Simulation results show that both of the proposed algorithms can eliminate the estimation errors introduced by fast correlated algorithm with a better estimation performance.