Research On Aid Decision-Making For Driving Safety Of Uncrewed Mining Trucks In Open-pit Mine

As the main production factor in the mining area,open-pit trucks' transportation has a high demand for safety.Uncrewed transportation can reduce maintenance costs and have specific adaptability to the dangerous or harsh working environment.However,open-pit mining areas have complicated geological structure,such as many detours and forks road with poor road stability.In the dust,fog,rain,snow,or other bad weather conditions,the perception of uncrewed mining trucks would be weakened.It may cause crashes,rushes off the road,etc.Therefore,how to ensure the safety and trust of road cars coordinated fusion control in extreme road condition,will be a challenge in studying uncrewed mining truck transportation.The intersection of the open-pit mine area has a high incidence of accidents due to diverse road conditions.Roads in the mining area are all semi-structured roads,lacking indicators like lane lines and arrows,without traffic lights and traffic signs.In order to ensure driving safety,it is very necessary to define safe passage rules.Using these rules to constrain uncrewed mining trucks' driving process and aided trucks in passing the intersection safely.At present,research on behavior decision-making of uncrewed mining trucks at unsignalized intersections mainly focuses on urban traffic scenarios.Meanwhile,research on uncrewed mining trucks driving in open-pit mines focuses on route planning,route avoidance,road rights allocation.However,there are few studies on the behavioral decision-making of uncrewed mining trucks.Therefore,the existing research on uncrewed mining truck decision-making in the open-pit mining environment have some shortcomings:Firstlty,The environmentally constrained safety distance model of uncrewed mining truck in open pit areas is missing.The open-pit mining environment has more factors affecting the driving safety of uncrewed mining truck,the existing safety distance model is unreasonable,the safety constraints are not complete,and it is difficult to adapt to the application requirements of open-pit mining environment.Secondly,Existing decision-making methods are difficult to meet security requirements in environmental change.With the vehicle travel,driving safety constraints change,the existing decision-making methods are difficult to adjust or respond to the changes adjusting driving safety constraints in a timely and reasonable manner,the decision-making is in violation of the current driving safety constraints,resulting in inefficient traffic,and even vehicle crashes.To sum up,this paper is based on the actual open pit mining environment,analyzing the influence of extreme environment characteristics of the mining area and intersection traffic characteristics on driving safety,and establishes the safety distance model with constraint for the possible rear-end collision and intersection conflict at the intersection respectively.Under the constraint of the environment,the continuous state decision problem is discretized,and use reinforcement learning to model the decision problem.Q-learning is used to solve the convergence Q table of the decision problem model,which can be regarded as the expression form of the behavior rule base obtained after training.Based on the roadside equipment and edge computing node that can realize the full coverage of the road in the mining area,the trained behavioral rule base can be deployed to the edge computing node to provide auxiliary decision-making uncrewed mining car to pass the intersection.In this paper,the study on the traffic aid decision of the driverless mining car under the accident-prone traffic intersections in open-pit mining areas is carried out to provide the basis for improving the work efficiency and ensuring the safety of transportation operations.The main work completed in this paper is as follows:Firstly,establish a safety distance model with environmental constraints.Unlike urban roads,the operating transportation environment in open-pit mining areas is unique,and the impact of extreme environments on driving safety needs to be considered when formulating safety rules.This paper analyzes the environment's influence on the driving safety of uncrewed mining vehicles based on the existing driving restrictions in the open-pit mining area and formulates fuzzy inference rules.The acceleration constraint factor fuzzy inference system is used to solve the maximum acceleration and deceleration constraint factors of uncrewed mining vehicles in different situations.Analyze the interaction process and intersection conflicts of the uncrewed mine workshop at road intersections in the open-pit mining area,and introduce the maximum acceleration and deceleration constraint factors for the rear-end conflicts and cross conflicts at the road intersections in the open-pit mining area,and establish a safe distance with environmental constraints model.Secondly,from the perspective of ensuring driving safety and improving transportation efficiency,based on the construction of driving safety constraint rules,an auxiliary decision-making method for the safe passage of uncrewed mining trucks in open-pit mining areas is proposed.The decision-making problem of uncrewed mining trucks passing through the open-pit mine intersection is regarded as a Markov decision-making problem.The decision-making problem is solved by the reinforcement learning method.During the solution process,the safety distance model with environmental constraints is used to solve the problem of unmanned mines.The vehicle status is classified,and the behavior of the unmanned mining vehicle is restrained accordingly.The Q table obtained by training convergence can be regarded as a behavior rule library after training to realize the auxiliary decision-making for unmanned mine vehicles' safe passage at intersections of open-pit mining areas.Finally,a simulation experiment platform was built to verify the acceleration constraint factor,the safety distance model with constraints,and the traffic aid decision-making method.Use Pre Scan 8.5.0 GUI to build an open-pit mine simulation environment and perform system modeling and simulation analysis under MATLAB&Simulink to verify the effectiveness of the maximum acceleration and deceleration constraint factor fuzzy inference system and the safety distance model with environmental constraints.The program of the decision algorithm is implemented in MATLAB,and the environment model is established by considering the road conditions at the intersection in the open-pit mining area.The Poisson distribution is used to fit the situation of uncrewed mining trucks driving out of the intersection.By comparing the average return received by the system under different decision-making scopes,traffic flow,and the delay time of a single vehicle,the decision-making algorithm's effectiveness in ensuring driving safety and improving transportation efficiency is verified.