Personal Trajectory Analysis Of Coal Mining

The underground personnel positioning system is one of the "six systems" of the underground safety and hazard avoidance system,which accumulates a lot of historical data in the long-term operation process.Through the analysis and mining of these data,people can find out the movement rules of personnel and improve the level of coal mine safety management.The trajectory information formed by the location data of underground coal mine personnel contains real-time space-time information of underground personnel,and contains personal behavior and state of underground personnel,such as individual’s type of work,operation activities,intention and so on.It contains the group behavior and state of underground personnel,such as team operation,shift handover and so on.In this paper,based on the trajectory data generated by the underground personnel positioning system of coal mine,using the trajectory data analysis method to mine the rules contained in the trajectory data of coal mine.The following aspects are mainly studied in this thesis.Firstly,in order to solve the problem of fast query for a large number of trajectory data,a GPU parallel processing framework suitable for large data background is proposed.According to the characteristics of the candidate set in Top-K query in actual operation,the GPU-based trajectory query of coal mine personnel is proposed called TTK_GPU.This method queries candidate sets of Top-K trajectories in GPU by parallel processing,which effectively ensures load balancing and improves query speed.Secondly,aiming at the difficulty of extracting frequent patterns from underground trajectory,an approximate frequent pattern discovery algorithm TFPMA based on prior conditions is proposed.Aiming at the problem of large amount of trajectory data,a divide-and-conquer processing method based on TFPMA algorithm,TFPMA_DC is proposed.It can not only maintain the memory requirement within the manageable range,but also realize parallelism,and achieves good results.Thirdly,in view of the nature of the classification problem of personnel role determination,based on the trajectory data of underground personnel,a method of determining the role of personnel based on the combination of ICNN and LSTM,DPRT,is proposed.ICNN is used to extract the depth characteristics of trajectory data,LSTM is used to realize the long-term dependence model of trajectory data,and the role of personnel is determined by training depth neural network.Finally,three main function modules are added to the system of personnel location and attendance in coal mine,which are fast inquiry of personnel trajectory under reference trajectory,acquisition and display of frequent trajectory of underground personnel,and identification of work types under the absence of job information,so as to enhance the function of the original system and provide software support for safety management.