Research On Multi-sensor Information Fusion And Intelligent Sensing Technology For Driverless Trackless Rubber-tyred Vehicles

In recent years,China has begun to accelerate the construction of intelligent coal mines,among which intelligent auxiliary transport system is one of the most important research and development contents.Trackless rubber-tyred vehicle is an important transportation equipment in the auxiliary transportation system of mine.Driverless driving is the intelligent development direction that must be passed in the future.At present,the research on the underground driverless technology of the trolley is still in the primary stage of structural design and theoretical exploration,and there is still a lack of key technology research on the intelligent perception of the trolley's driverless driving in the special environment of the underground roadway.Therefore,it is of great theoretical significance and practical value to carry out the research on intelligent sensing technology of the trackless rubber-tyred vehicle for promoting the development of the trackless rubber-tyred vehicle driverless technology and accelerating the realization of the development goal of “It's safest when nobody's around “ in the mine.Based on the method of multi-sensor information fusion,this paper studies the intelligent perception technology of the driverless trackless rubber-tyred vehicle from two aspects of the hardware of the perception system and the information fusion strategy,based on the special working conditions of the underground roadway and the prototype of the coal mine passenger trackless rubber-tyred vehicle.First of all,through the analysis of the downhole trailless rubber-tyred vehicle perception condition and perceived demand,the driverless driving perception system is designed.Based on the sensing system as the hardware foundation,the demand analysis of multi-sensor information fusion is carried out,and a two-level information fusion strategy of “target level + decision level” is proposed.Secondly,a target-level information fusion strategy based on YOLOV4 algorithm was designed,and road simulation tests of target-level fusion perception were carried out in daytime and night driving environments respectively.The results show that the target-level information fusion algorithm proposed in this paper can correlate the ROI regions generated by various sensors,and reduce the rate of false detection and missed detection of millimeter wave radar and machine vision algorithm.Thirdly,by improving the D-S evidence theory of the existing decision-level information fusion method,a decision-level information fusion strategy based on the improved D-S evidence theory algorithm was designed,and a decision-level fusion sensing road simulation test was carried out.The results show that decision-level fusion can reduce the impact of conflicting evidence in the results of the target level fusion,further improve the accuracy of decision and reduce the probability of false detection.Finally,facing the working conditions of the trolley and its underground environment,the simulated test vehicle and the simulated tunnel environment were selected to carry out the tunnel simulation test of the environment perception of the trolley.The results show that the two-level fusion sensing method can make the sensor performance complementary,improve the perception accuracy and reduce the rate of missed detection.The multi-sensing information fusion intelligent perception system designed in this paper,which is based on the two-level fusion strategy of “target level + decision level”,can meet the accuracy requirements of environment perception for the underground driverless trackless rubber-tyred,and can provide theoretical and technical support for the development and application of the driverless trackless rubber-tyred vehicle.There are 51 figures,23 tables and 120 references in this paper.