Design And Research On Automatic Driving System Of Explosion-Proof Battery-Type Electric Locomotive

At present,China ’s coal mining has gradually moved from mechanization and automation to intelligence.The development goals of intelligent mines and unmanned mines have become the industry consensus.In order to solve problems of high accident rate and low transportation efficiency of manual driving electric locomotive,this thesis takes CTL 12 / 6 G P explosion-proof battery electric locomotive as the research object,and designs automatic driving control system,including design of On-board control system,development of underground turnout and signal light control system,UWB positioning non-line-of-sight error compensation algorithm research,and mine roadway image target detection algorithm research based on Faster RCNN,which provides technical support for intelligent driving of electric locomotive.In order to solve problems of complicated driving information and low intelligence of control system,the On-board control system is designed.According to the overall scheme and driving conditions,the system function structure and the selection and installation of each module equipment are improved,and then the control system based on PLC is designed.In order to solve problems of high cost,low efficiency and poor safety of manual switching turnout,the control system of underground turnout and signal lamp is designed.Combined with the ad hoc network base station network,the hardware architecture of the system is improved and the layout of the substation acquisition board is optimized.Considering the safety of electric locomotive operation,the operation logic of control program is designed.In order to solve the problems that the dusty and complex environment of coal mine roadway leads to the increase of UWB error,the research on the non-line-of-sight error compensation algorithm of UWB positioning is carried out.By extracting the characteristics of the impulse response waveform channel of the data set,the integrated learning method based on Ada Boost is used to identify the propagation channel of LOS/ NLOS,and then combined with the UWB ranging noise model,the EKF algorithm is used to compensate the UWB non-line-of-sight error.Finally,through experiments,the application effect of LOS / NLOS identification algorithm and the positioning accuracy and stability of EKF error compensation algorithm in different NLOS noise environments are verified.In order to solve the problems of slow detection speed and poor accuracy caused by complicated types and large-scale differences of roadway equipment,the research on roadway image target detection algorithm based on Faster RCNN is carried out.By comparing the noise reduction effect of filtering algorithm and the enhancement effect of Retinex algorithm to improve the quality of roadway image.The roadway image target detection model based on Faster RCNN is constructed,which is divided into six detection categories: track,personnel,pipeline,equipment,train and coal block.The data set is expanded by data augmentation,and the hyperparameters are optimized by200 iterations with the help of Pytorch framework.Finally,the detection effect of the model is verified by image and video experiments.In this thesis,automatic driving technology is applied to the field of auxiliary transportation in underground coal mines.By using information and control technology to improve the level of transportation automation,the safety and efficiency of transportation operations can be improved,which can bring considerable economic benefits to coal mine production enterprises.This thesis has 55 figures,20 tables and 84 references.