Research On Application Of Compressed Sensing In Mine Internet Of Things

Since the concept of the Internet of Things was introduced,it has been applied to various industrial productions,and the Mine Internet of Things is one of them.With the continuous development of social technology,more and more wireless network nodes are distributed in the underground of coal mines and on the ground,which causes many problems to be solved in the mine Internet of Things,and because of the great danger in underground work,many Coal mining companies began to focus on unmanned operations.But this also means that the information that needs to be collected and processed will grow geometrically,which will greatly limit the computing,storage and endurance of nodes in the Internet of Things.Compressed sensing theory,as a new sampling theory,has attracted the attention of many scholars in recent years.The sampling frequency of the sampling algorithm is much lower than the traditional Nyquist sampling frequency,which can reduce the sampling amount of data,reduce the resource loss of the Internet of Things when it is used,and also improve the data transmission and storage capacity.Based on the analysis of traditional compressed sensing theory,this thesis focuses on the optimization of compressed sensing algorithm,and proves that it can be better applied to mine Internet of Things and make signal compression more efficient through experimental simulation.First,we will detail three important aspects of the theory of compressed sensing,namely sparse representation,observation matrix and restoration reconstruction,as well as theoretical algorithms often involved in each link.Then,based on this,the algorithm is optimized for sparse representation and restoration reconstruction.The first spark-level sparsity and the tail minimization,the algorithm has a measure theoretical uniqueness for nearly sparse-level sparsity from compressed measurements.Specifically,it is assumed that the m rows of the matrix A have a Spark property,and it is assumed that the sparsity s is larger than m/2 and smaller than m.Then the solution of the compressed measurement is unique.On each sparse plane,there can be a set of measures of zero.This phenomenon was observed and confirmed by the L₁ tail miniaturization procedure,which uniquely restored the sparse signal in thousands of random experiments,and s is greater than m/2.The second is the combined feedback fast recovery threshold algorithm in zero space.The algorithm is based on the framework of zero space adjustment,and the sparse operator is added to the framework to obtain the algorithm under different conditions.Finally,using Matlab2017a as the simulation platform,the optimization algorithm is combined with the relevant data of the coal mine,and the m file is written to call the image data to carry out the experiment and simulation of the algorithm.At the same time,the optimization algorithm is compared with the original compressed sensing algorithm to prove the feasibility of the algorithm in the mine Internet of Things.Figure[32]Table[1]Reference[64]...