Research On The Real-Time Coding Of Distributed Sources In Coal Mines

Mine Internet of Things is the product of coal mine informationization,and it is an important technical means to realize the all-round perception of coal mine safety production.The promotion of mine Internet of Things is of great significance to improve the safety production management level of coal mines.In the mine Internet of Things,the amounts of data generated by many sensors are huge,which has a huge impact on the data transmission network.Therefore,how to compress and transmit the real-time data generated by the sensor is a key issue worthy of study in the mine Internet of Things.In this thesis,the following research is carried out on the real-time data compression problem generated by coal mine distributed sources:(1)The data generated by the sensors are huge and the data change slowly.Meanwhile the traditional sampling method can not capture the sudden change of the signal in the sampling gap.This thesis uses distributed compressed sensing to process the real-time data collected by the sensors.Distributed Compressed Sensing is a new information acquisition method that breaks through the Nyquist sampling theorem.It can capture the abnormal change information of the signal in time,and can simultaneously utilize the correlation between the signal and the signal,so as to effectively reduce the amount of data transmitted by the network and especially be suitable for distributed application scenarios of the mine Internet of Things.(2)Taking the typical source-gas concentration source of coal mine as an example,the characteristics of state,spatial correlation,sparsity and joint sparsity are analyzed.It is verified that the gas concentration signal is sparse and joint-sparse on the FFT basis.Sparseness proves that the gas concentration signal can be processed by the distributed compressed sensing theory,and the effect of distributed compressed sensing is better than that of compressed sensing when the observed values are the same.(3)Analyze the advantages and disadvantages of Gaussian random observation matrix,Bernoulli random observation matrix and sparse random observation matrix commonly used in the research of distributed compressed sensing theory,and then an optimized sparse random matrix is proposed according to the low complexity requirement for the real-time data processing algorithm of sensor nodes in mine Internet of Things.After the RIP characteristics of the optimized sparse random matrix are proved,the performance analysis and verification are carried out.Theresults show that the optimized sparse random matrix has mutual coefficient with the FFT base and calculation complexity and observation effects are better than other matrices selected.(4)Based on the theory of distributed compressed sensing,a real-time variable rate coding strategy for distributed source in coal mine is proposed.Then the virtual clustering method in the coding strategy is introduced.Finally,a compression algorithm based on the run-length coding idea is proposed for the slowly changing data in the distributed source of coal mine.The experimental results show that the proposed real-time coding strategy can be used very well.The real-time data generated by the distributed source of the coal mine is compressed better.