Applications And Researches On Distributed Process-based Sources Coding

The application of mines Internet of Things(IoT)requires deploying a large number of wireless sensors.However,the rapid increase in the number of wireless sensors will bring about such problems as the sharp increase of network data and the decline of transmission reliability.At the same time,mining wireless sensor nodes are also facing energy-constrained problems.Many studies have shown that as long as the amount of network data is reduced,these problems can be effectively solved.Therefore,the research of efficient and reliable coding and decoding methods has become a hot spot in the development of mine IoT.The distributed compressed sensing(DCS)can realize signal coding with lower sampling rate and reduce the data capacity transferred in the network effectively,so it is very suitable for mine IoT scenario.In this thesis,coal mine microseismic signals are taken as an example to analyze the spatial correlation and sparsity.And we compare the codec effects of the compressed sensing(CS)and DCS through experiments.The experimental results show that DCS has better performance in the codec of microseismic signals due to the utilization of intra-signal correlation and inter-signal correlation.The random measurement matrix has the disadvantages of high coding complexity and is not conducive to hardware implementation,so an LDPC-like matrix based on low-density parity-check code(LDPC)is studied in this thesis.We analyze the influence of row weight of LDPC measurement matrix on the measurement accuracy of microseismic signals,and construct an LDPC-like measurement matrix suitable for coal mine microseismic signals.Experiments show that when the sampling rate is greater than 0.35,the LDPC-like observation matrix(row weight is 20)has a good recovery effect.A bat-inspired decoding algorithm for CS(CS_BA)is proposed by combining the Bat Algorithm(BA)and the pruning technique in subspace pursuit,and we extend the algorithm to the joint decoding of DCS.Experiments show that compared with the traditional greedy pursuit algorithm,CS_BA algorithm has better recovery ability without channel noise.For mine microseismic signals,the average recovery error of the bat-inspired joint decoding algorithm for DCS(DCS_BA)is 0.1 lower than the CS_BA algorithm.However,the anti-noise capability of CS_BA algorithm is poor,the recovery accuracy of CS_BA algorithm decreases greatly in the case of large noise.Aiming at the problem of poor anti-noise capability of greedy pursuit algorithm,a decoding algorithm based on layered belief propagation(CS_LBP)is proposed.And we extend the algorithm to DCS,and further propose a layered belief propagation joint decoding algorithm based on side information(DCS_LBP).Experiments show that the coal mine microseismic signal can be recovered well by using CS_LBP algorithm and DCS_LBP algorithm in the case of adding channel noise.In particular,the DCS_LBP algorithm can recover the microseismic signal better by using the side information at a very low sampling rate.