Research On Equalization Technology For Sparse Multipath Channel

The multipath and fading characteristics of sparse multipath channels have influences on the receiver performance of communication systems,result in severe inter-symbol interference of received signals.In order to increase the quality of the signals received,equalization technology is employed at the receiving end to compensate the distortion caused by the channel.The existing equalization technology is unable to be a fast and effective compensation method for deep fading sparse multipath channels,as well as unsuitable for burst communication systems exampled as underwater acoustic communication.Thus,we start from the structure of equalizer,the characteristics of equalizer coefficients and the design of equalization algorithm to enhance the performance and practicability of equalizer for sparse multipath channels.The main work and achievements proposed in this paper are summarized as follows:Firstly,for better convergence performance of equalizer in deep fading sparse multipath channels at low signal-to-noise ratio,a modified adaptive sparse equalization algorithm based on channel shortening is proposed.Using short training sequence,the front-end equalizer is designed according to minimum mean square error criterion,partially effectively compensating the deep fading distortion of the original channel.The back-end equalizer adopts stochastic gradient pursuit algorithm to update coefficients,eliminating the residual inter-symbol interference of the equivalent channel.Compared with single-stage fractional-spaced adaptive equalizer,the novel algorithm provides superior performance by increasing the convergence rate and decreasing the computation.Secondly,for fast convergence rate of blind equalizer in sparse multipath channels,theoretical inference has proved that the equalizer for sparse multipath channels is sparse.According to the constant modulus characteristics of MPSK signal,by adding a sparsity-inducing ₀l-norm penalty on the equalizer tap coefficients,two fast convergence blind equalization algorithms for sparse multipath channel based on sparse adaptive filter are proposed.The algorithm 1 not only assigns step sizes according to the magnitude of the equalizer tap weights,resulting in large step sizes for the active taps,small step sizes for the inactive taps,but also making adaptive zero adjustment for the inactive taps;The algorithm 2 takes the advantages of the recursive least square algorithm,as well as attracting the inactive taps to zero.The novel algorithms outperform the existing blind equalization algorithms for sparse channel in convergence rate at the same residual inter-symbol interference level.Thirdly,in order to improve the blind equalization performance for deep fading sparse multipath channels,a ₀l-norm constraint fractionally spaced sparse adaptive dual-mode algorithm is proposed.Based on the traditional fractionally spaced dual-mode blind equalization algorithm,and motivated by the sparse adaptive filter theory,by appending constraint sparsity on the equalizer coefficients with a l₀-norm penalty,the algorithm assigns step sizes proportionate to the magnitude of the current individual tap weights to corresponding taps as well as attracting the inactive taps to zero adaptively.Compared with the traditional algorithms,the novel algorithm outperforms in increasing the convergence rate and decreasing the residual inter-symbol interference.Finally,by applying the above blind equalization technology for sparse multipath channels,novel blind equalization and blind synchronization algorithm is presented for blind demodulation of underwater acoustic burst-mode signals.The blind equalization of burst-mode signals is accomplished by a fractionally-spaced equalizer based on sparse adaptive filter and data-reusing.Then using nonlinear transformation and the second-order digital phase-locked loop,the output of the equalizer can estimate the carrier frequency and phase offset.Experiment results show the effectiveness of the novel algorithm to blind demodulation of burst-mode signals,in shallow underwater acoustic multipath channels simulated by the BELLHOP model.