Research And Application Of The Compressed Sensing Reconstruction Algorithms Based On Dice Coefficient

With the booming of the technology era,the demand for massive data is increasing.If the Nyquist sampling criteria continues to be used,must require that the sampling frequency be no less than twice the maximum frequency of the signal to ensure reconstruction accuracy.Therefore,the sampled data will contain a large amount of redundant information,resulting in waste of resources and putting tremendous pressure on the transmission and storage of information,The emergence of compressed sensing(CS)theory breaks the bottleneck of the traditional Nyquist sampling theorem.It is a new sampling theorem mode and its core idea is that sampling and compression are carried out simultaneously,which is destined to be an innovation in the field of information processing and communication.Channel estimation is one of the key technologies in wireless communication systems.The theoretical premise of compressed sensing is that the signal is sparse,and the impulse response of most wireless multipath channels are also sparse.This sparse feature undoubtedly provides reliable theoretical basis and broad research prospects for compressed sensing in the field of wireless multipath channel estimation.Firstly,in this thesis,the principle of compressed sensing theory is studied in detail.Secondly,the core principles and performance simulation and analysis of three traditional compressed sensing greedy reconstruction algorithms are described.Then,in view of there are some shortcomings in the traditional compressed sensing reconstruction algorithms for matching the best atoms,so the atom matching criterion based on Dice coefficient is studied to optimize the traditional greedy algorithms.Finally,combined with the sparse characteristics of wireless multipath channel,the sparse channel estimation of compressed sensing based on Dice coefficient is studied.The main work of this thesis has the following three points:(1)Three traditional greedy algorithms are studied:the matching pursuit(MP)algorithm,orthogonal matching pursuit(OMP)algorithm and stagewise weak orthogonal matching pursuit(SWOMP)algorithm.The core principles of the three reconstruction algorithms are analyzed in detail,and the reconstruction performance is simulated and analyzed.(2)In view of the problems existing in the traditional compressed sensing reconstruction algorithms,such as the method of selecting the best atoms by using the inner product criterion does not preserve the original state of the vector well,and it is difficult to distinguish between similar atoms,so resulting in inaccurate matching and the important information of the original signal can not better highlight This thesis studies the matching criterion of Dice coefficient instead of the traditional inner product criterion.The criterion can optimize the support set,locate the main components of the residual quickly,reduce the influence of matching results between any two similar atoms and residual signals,and improve the reconstruction performance.Therefore,this thesis designs DMP,DOMP and DSWOMP algorithms based on Dice coefficient matching criterion.In order to further verify the effectiveness of these algorithms,these algorithms are used to simulate and reconstruct the one-dimensional time domain signal,and the performance indexes such as the residual vector's modulus,reconstructed success rate and reconstructed time are simulated and compared.The simulation results show that compared with the traditional greedy algorithms based on inner product criterion,the greedy algorithms based on Dice coefficient matching criterion have better performance(3)Compared with the traditional channel estimation algorithms,the algorithms that combined of compressed sensing and channel estimation have important advantages:it can improve the utilization of system resources and the accuracy of channel estimation.Therefore,combined with the theory of compressed sensing,the sparse multipath channel estimation algorithms based on Dice coefficient are proposed and compared with the traditional least squares.The simulation results show that compared with the traditional algorithms,these algorithms based on Dice coefficient matching criterion have higher estimation accuracy in sparse channel estimation and can obtain better channel estimation performance.In summary,these algorithms designed in this thesis have higher reconstruction accuracy and improve the performance of sparse channel estimation.It provides a new solution for channel estimation in the field of communication and has theoretical research value.