Multi-scale variable-domain sensing-control cooperative control system and method for trackless rubber-tyred vehicle

Abstract

The invention discloses a trackless rubber-tyred vehicle multi-scale variable-domain sensing-control cooperative control system and method.The system comprises an environment sensing module, a long-acting self-adaptive decision module, a local intelligent aging control module and a cooperative control module, and the environment sensing module is used for obtaining surrounding environment information; the cooperative control module is used for fusing the outputs of the long-acting self-adaptive decision module and the local intelligent aging control module, and introducing cooperative control weight parameters to dynamically allocate control weights between the long-acting self-adaptive decision module and the local intelligent aging control module; by setting a weight formula and based on a dynamic weight optimization method of DDPG, intelligent dynamic adjustment of cooperative control weight parameters is realized; and finally, local trajectory planning and trajectory tracking are performed according to the current local obstacle information while global path planning is performed on the trackless rubber-tyred vehicle, so that the running trajectory of the trackless rubber-tyred vehicle is adjusted and controlled, the smoothness of the running trajectory is realized, and higher stability and safety in the running process of the trackless rubber-tyred vehicle are ensured.

Description

Technical Field

[0001] The present invention relates to a multi-scale variable-domain sensing and control collaborative control system and method for a trackless rubber-tyred vehicle, belonging to the technical field of intelligent driving of coal mine vehicles. Background Art

[0002] Rubber-tyred trackless vehicles are rubber-wheeled transport vehicles designed to operate on the floor of underground tunnels. Composed of a tractor and a carrier, these vehicles are used to transport personnel, materials, small equipment, sand, and other materials, in addition to coal. Due to their flexible scheduling and ability to transport long, continuous distances, they are currently the primary means of transportation in coal mines.

[0003] As coal mines continue to advance in mechanization, automation, intelligence, and informatization, the intelligence level of rubber-tyred trackless vehicles in coal mines is also increasing. However, due to the complex underground environment and narrow tunnels, the driver's vision is sometimes obstructed, making it difficult to accurately determine whether obstacles are present in all directions that could impact vehicle and production safety, hindering the safety of life and property of mine personnel. Consequently, the trend toward autonomous driving is underway. However, current autonomous vehicle control often relies on pre-set, fixed models, while autonomous time-sensitive control requires real-time adaptation to changes in the tunnel environment (i.e., local adjustments to the route based on detected obstacles during operation, which deviates from the overall path plan). This leads to a conflict between the long-term trajectory adaptive decision-making of the trackless vehicle and the time-sensitive control of the autonomous vehicle. Because long-term decision-making focuses on stability and overall efficiency, it may overlook immediate local adjustments. Meanwhile, time-sensitive control, with its overly rapid local response, can disrupt the overall plan and affect vehicle operational stability. If the conflict between the two cannot be effectively resolved, the vehicle may not be able to ensure continuous optimization of its trajectory when facing complex environments, and may also lose the necessary emergency response capabilities, seriously affecting the stability of the coal mine rubber-tyred trackless vehicle and the safety of mine operations.

[0004] Therefore, how to provide a new system and method that can automatically adjust the weight between long-term decision-making and unmanned driving time control according to the surrounding environment during driving, so as to effectively ensure the operation stability and working safety of trackless rubber-tyred vehicles, is the research direction required by the present invention. Summary of the Invention

[0005] In response to the problems existing in the above-mentioned prior art, the present invention provides a multi-scale variable-domain sensing and control collaborative control system and method for a trackless rubber-tyred vehicle, which can automatically adjust the weight between long-term decision-making and unmanned driving time control according to the surrounding environmental conditions during driving, thereby ensuring continuous optimization of the overall trajectory during driving and having emergency response capabilities, thereby achieving stable operation and driving safety of the trackless rubber-tyred vehicle.

[0006] In order to achieve the above-mentioned object, the technical solution adopted by the present invention is: a multi-scale variable domain sensing and control collaborative control system for a trackless rubber-tyred vehicle, comprising an environmental perception module, a long-term adaptive decision module, a local intelligent time-effect control module and a collaborative control module arranged on the trackless rubber-tyred vehicle; The environmental perception module is used to detect the surrounding environment information of the trackless rubber-tyred vehicle, including the lane structure and obstacle information; The long-term adaptive decision-making module is used to plan the global operation path of the trackless rubber-tyred vehicle according to the trackless rubber-tyred vehicle operation lane map; The local intelligent time-effect control module is used to obtain the surrounding environment information fed back by the environmental perception module in real time, and plan the local path of the trackless rubber-tyred vehicle and realize adaptive trajectory tracking control based on the trackless rubber-tyred vehicle's driving trajectory and speed; The collaborative control module is used to integrate the long-term adaptive decision-making module and the unmanned driving time control module, and introduce collaborative control weight parameters to dynamically allocate the control weights between the two, thereby controlling the running trajectory of the trackless rubber-tyred vehicle.

[0007] Furthermore, the environmental perception module includes a lidar, a 4D millimeter-wave radar, a visible light camera and an infrared camera; the visible light camera and the infrared camera are used to identify obstacles on the driving path and classify the obstacles; the lidar and the 4D millimeter-wave radar are used to obtain information on the lane structure, obstacle height and driving speed.

[0008] The control method of the multi-scale variable-domain sensing and control collaborative control system for the above-mentioned trackless rubber-tyred vehicle includes the following steps: Step 1: The trackless rubber-tyred vehicle uses the laser radar, visible light camera, and infrared camera in the environmental perception module to obtain information about the surrounding environment. The laser radar and 4D millimeter-wave radar obtain information about the lane structure, obstacle height, and speed, and build a three-dimensional high-precision point cloud map. The visible light camera and infrared camera identify obstacles on the driving path and classify them. Step 2: Based on the operational characteristics of rubber-tyred trackless vehicles, the operation process of rubber-tyred trackless vehicles in the roadway is divided into long-term autonomous decision-making of the trackless rubber-tyred vehicle in a large-scale space and short-term local time-sensitive control decision-making to deal with special situations. The long-term adaptive decision-making module and the local intelligent time-sensitive control module are used for planning respectively; Step 3: Based on the prior map of the underground coal mine tunnel, a long-term adaptive decision-making module for trackless rubber-tyred vehicles is established according to the genetic algorithm to realize the global path planning and control of the trackless rubber-tyred vehicles; Step 4: Establish a local intelligent time-sensitive control module through the model predictive control algorithm. Based on the historical trajectory information and environmental information of the trackless rubber-tyred vehicle, it is used to adjust the control parameters in a short time to deal with special situations, improve the control flexibility and adaptability of the trackless rubber-tyred vehicle, and realize the local path planning control of the trackless rubber-tyred vehicle; Step 5: Establish a collaborative control module, integrate the outputs of the long-term adaptive decision-making module and the local intelligent time-sensitive control module, and introduce a collaborative control weight parameter to dynamically allocate the control weight between the two; Step 6: The environmental perception module evaluates the environment surrounding the trackless rubber-tyred vehicle. Based on the dynamic weight optimization method of DDPG, the collaborative control weight parameters are intelligently and dynamically adjusted by adaptively learning the changes in the environment and vehicle status. Step 7: Based on the weight ratio of the long-term adaptive decision-making module and the local intelligent time-efficiency control module output by the collaborative control module in real time in step 6, while performing global path planning, local trajectory planning and trajectory tracking are performed based on the current local obstacle information, thereby adjusting and controlling the running trajectory of the rubber-tyred trackless vehicle to achieve smooth running trajectory.

[0009] Furthermore, the step three of establishing a long-term adaptive decision-making module for trackless rubber-tyred vehicles based on a genetic algorithm is specifically as follows: The genetic algorithm continuously optimizes the path through selection → crossover → mutation, and finally finds the optimal path for global path planning of trackless rubber-tyred vehicles. The specific process is as follows: Initialize the population to randomly generate N feasible paths, each path P is represented by a coordinate sequence: , Calculate the comprehensive score for each path P: in ; Select Operation P: Crossover: Randomly select parents and , swap the fragments at position k; Mutation: Waypoints Random perturbations: ; New generation of population: final It terminates when it converges or reaches the maximum number of iterations, and generates the optimal path: .

[0010] Furthermore, the specific formula of the model predictive control algorithm in step 4 is: in is the running track state of the rubber-tyred trackless vehicle at the next moment, is the current state, Indicates the current speed and steering angle of the rubber-tyred trackless vehicle.

[0011] Furthermore, the collaborative control weight parameter in step 5 is α, and the weight calculation is performed using this parameter. The specific formula is: Where: and are the outputs of the long-term adaptive decision-making module and the local intelligent time-sensitive control module respectively, α is the collaborative control weight parameter, and 0≤α≤1.

[0012] Furthermore, the dynamic weight optimization method of DDPG in step 6 is specifically as follows: Initialize, select action Execute an action , the reward function is , the new state is ; Storage experience arrive ; Sampling small batches - ; Update Critic: minimize Update Actors: Soft update TargetNetworks: Where s represents the current state of the rubber-tyred trackless vehicle, which is obtained by the surrounding environment information provided by the environment perception module; is the output action of the weight optimizer, which is the value of the collaborative control weight parameter α; r is the reward function, which represents the deviation between the actual trajectory and the planned trajectory and the smoothness of the trackless rubber-wheeled vehicle and the reverse trajectory during operation. The smaller the deviation and the smoother the running trajectory, the greater the reward; the weight optimizer selects the operation with the largest Q value according to the real-time environmental information fed back by the environmental perception module, and obtains the corresponding The value is the optimal value of the collaborative control weight parameter α.

[0013] Compared with the prior art, in order to solve the problem of conflict between the long-term adaptive decision-making and unmanned driving time control of trackless rubber-tyred vehicles, the present invention establishes a collaborative control module, utilizes the output of the integrated long-term adaptive decision-making module and the local intelligent time control module, and introduces collaborative control weight parameters to dynamically allocate the control weights between the two; by setting the weight formula and the dynamic weight optimization method based on DDPG, the collaborative control weight parameters are intelligently and dynamically adjusted through the changes in the adaptive learning environment and vehicle state; and then the output weight ratio of the long-term adaptive decision-making module and the local intelligent time control module is determined in real time. Finally, while the trackless rubber-tyred vehicle is performing global path planning, it also performs local trajectory planning and trajectory tracking according to the current local obstacle information, thereby adjusting and controlling the running trajectory of the trackless rubber-tyred vehicle, achieving the smoothness of the running trajectory, and ensuring that the trackless rubber-tyred vehicle has higher stability and safety during operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 It is an overall flow chart of the present invention. DETAILED DESCRIPTION

[0015] The present invention will be further described below.

[0016] A multi-scale variable-domain sensing and control collaborative control system for a rubber-tyred trackless vehicle, comprising an environmental perception module, a long-term adaptive decision-making module, a local intelligent time-effect control module, and a collaborative control module, all of which are arranged on the vehicle. The environmental perception module is used to detect the surrounding environment information of the trackless rubber-tyred vehicle, including lane structure and obstacle information. The environmental perception module consists of four 4D millimeter-wave radars, one laser radar, one visible light camera and one infrared camera. The 4D millimeter-wave radars are installed at the four corners of the trackless rubber-tyred vehicle. The laser radar is higher than the 4D millimeter-wave radar and lower than the visible light camera and infrared camera. The visible light camera and infrared camera are used to identify obstacles on the driving path and classify them. The laser radar and 4D millimeter-wave radar are used to obtain lane structure, obstacle height and driving speed information. The long-term adaptive decision-making module is used to plan the global operation path of the trackless rubber-tyred vehicle according to the trackless rubber-tyred vehicle operation lane map; The local intelligent time-effect control module is used to obtain the surrounding environment information fed back by the environmental perception module in real time, and plan the local path of the trackless rubber-tyred vehicle and realize adaptive trajectory tracking control based on the trackless rubber-tyred vehicle's driving trajectory and speed; The collaborative control module is used to integrate the long-term adaptive decision-making module and the unmanned driving time control module, and introduce collaborative control weight parameters to dynamically allocate the control weights between the two, thereby controlling the running trajectory of the trackless rubber-tyred vehicle.

[0017] like Figure 1 As shown, the control method of the multi-scale variable domain sensing and control collaborative control system of the above-mentioned trackless rubber-tyred vehicle includes the following steps: Step 1: The trackless rubber-tyred vehicle uses the laser radar, visible light camera, and infrared camera in the environmental perception module to obtain information about the surrounding environment. The laser radar and 4D millimeter-wave radar obtain information about the lane structure, obstacle height, and speed, and build a three-dimensional high-precision point cloud map. The visible light camera and infrared camera identify obstacles on the driving path and classify them. Step 2: Based on the operational characteristics of rubber-tyred trackless vehicles, the operation process of rubber-tyred trackless vehicles in the roadway is divided into long-term autonomous decision-making of the trackless rubber-tyred vehicle in a large-scale space and short-term local time-sensitive control decision-making to deal with special situations. The long-term adaptive decision-making module and the local intelligent time-sensitive control module are used for planning respectively; Step 3: Based on the prior map of the underground coal mine tunnel, a long-term adaptive decision-making module for trackless rubber-tyred vehicles is established according to the genetic algorithm to realize the global path planning and control of the trackless rubber-tyred vehicles. Specifically: The genetic algorithm continuously optimizes the path through selection → crossover → mutation, and ultimately finds a short, safe, and smooth route as the optimal path for global path planning of trackless rubber-tyred vehicles. The specific process is as follows: Initialize the population to randomly generate N feasible paths, each path P is represented by a coordinate sequence: , Calculate the comprehensive score for each path P: in ; Select Operation P: Crossover: Randomly select parents and , swap the fragments at position k; Mutation: Waypoints Random perturbations: ; New generation of population: final It terminates when it converges or reaches the maximum number of iterations, and generates the optimal path: .

[0018] Step 4: Establish a local intelligent time-sensitive control module through the model predictive control algorithm. Based on the historical trajectory information and environmental information of the trackless rubber-tyred vehicle, the module is used to adjust the control parameters in a short time to deal with special situations, improve the control flexibility and adaptability of the trackless rubber-tyred vehicle, and realize the local path planning control of the trackless rubber-tyred vehicle. The specific formula of the model predictive control algorithm is: in is the running track state of the rubber-tyred trackless vehicle at the next moment, is the current state, Indicates the current speed and steering angle of the rubber-tyred trackless vehicle.

[0019] Step 5: Establish a collaborative control module, integrate the outputs of the long-term adaptive decision-making module and the local intelligent time-limited control module, and introduce a collaborative control weight parameter to dynamically allocate the control weight between the two. The collaborative control weight parameter is α, and the weight calculation is performed using this parameter. The specific formula is: Where: and are the outputs of the long-term adaptive decision-making module and the local intelligent time-sensitive control module respectively, α is the collaborative control weight parameter, and 0≤α≤1.

[0020] Step 6: The environment perception module evaluates the surrounding environment of the rubber-tyred trackless vehicle. Based on the dynamic weight optimization method of DDPG, the dynamic adjustment of the collaborative control weight parameters is realized by adaptively learning the changes in the environment and vehicle status. The dynamic weight optimization method of DDPG is specifically as follows: Initialize, select action Execute an action , the reward function is , the new state is ; Storage experience arrive ; Sampling small batches - ; Update Critic: minimize Update Actors: Soft Update Target Networks: Where s represents the current state of the rubber-tyred trackless vehicle, which is obtained by the surrounding environment information provided by the environmental perception module. The state specifically includes: the position and speed of the trackless rubber-tyred vehicle during operation, as well as the position and height information of surrounding obstacles; is the output action of the weight optimizer, which is the value of the collaborative control weight parameter α; r is the reward function, which represents the deviation between the actual trajectory and the planned trajectory and the smoothness of the trackless rubber-wheeled vehicle and the reverse trajectory during operation. The smaller the deviation and the smoother the running trajectory, the greater the reward; the weight optimizer selects the operation with the largest Q value according to the real-time environmental information fed back by the environmental perception module, and obtains the corresponding The value is the optimal value of the collaborative control weight parameter α; Step 7: Based on the weight ratio of the long-term adaptive decision-making module and the local intelligent time-efficiency control module output by the collaborative control module in real time in step 6, while performing global path planning, local trajectory planning and trajectory tracking are performed based on the current local obstacle information, thereby adjusting and controlling the running trajectory of the rubber-tyred trackless vehicle to achieve smoothness of the running trajectory and ensure higher stability and safety during the operation of the rubber-tyred trackless vehicle.

[0021] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A multi-scale variable domain sensing and control collaborative control system for a rubber-tyred trackless vehicle, characterized in that: It includes an environmental perception module, a long-term adaptive decision-making module, a local intelligent time-effect control module, and a collaborative control module deployed on the rubber-tyred trackless vehicle; The environmental perception module is used to detect the environmental information surrounding the trackless rubber-tyred vehicle; The long-term adaptive decision-making module is used to plan the global operation path of the trackless rubber-tyred vehicle according to the trackless rubber-tyred vehicle operation lane map; The local intelligent time-effect control module is used to obtain the surrounding environment information fed back by the environmental perception module in real time, and plan the local path of the trackless rubber-tyred vehicle and realize adaptive trajectory tracking control based on the trackless rubber-tyred vehicle's driving trajectory and speed; The collaborative control module is used to integrate the long-term adaptive decision-making module and the unmanned driving time control module, and introduce collaborative control weight parameters to dynamically allocate the control weights between the two, thereby controlling the running trajectory of the trackless rubber-tyred vehicle.

2. The multi-scale variable domain sensing and control collaborative control system for rubber-tyred trackless vehicles according to claim 1 is characterized in that: The environmental perception module includes a laser radar, a 4D millimeter-wave radar, a visible light camera, and an infrared camera; the visible light camera and the infrared camera are used to identify and classify obstacles on the driving path; the laser radar and the 4D millimeter-wave radar are used to obtain information on the lane structure, obstacle height, and driving speed.

3. A control method for a multi-scale variable domain sensing and control collaborative control system for a trackless rubber-tyred vehicle according to claim 1 or 2, characterized in that: The following steps are involved: Step 1: The trackless rubber-tyred vehicle uses the laser radar, visible light camera, and infrared camera in the environmental perception module to obtain information about the surrounding environment. The laser radar and 4D millimeter-wave radar obtain information about the lane structure, obstacle height, and speed, and build a three-dimensional high-precision point cloud map. The visible light camera and infrared camera identify obstacles on the driving path and classify them. Step 2: Based on the operational characteristics of rubber-tyred trackless vehicles, the operation process of rubber-tyred trackless vehicles in the roadway is divided into long-term autonomous decision-making of the trackless rubber-tyred vehicle in a large-scale space and short-term local time-sensitive control decision-making to deal with special situations. The long-term adaptive decision-making module and the local intelligent time-sensitive control module are used for planning respectively; Step 3: Based on the prior map of the underground coal mine tunnel, a long-term adaptive decision-making module for trackless rubber-tyred vehicles is established according to the genetic algorithm to realize the global path planning and control of the trackless rubber-tyred vehicles; Step 4: Establish a local intelligent time-sensitive control module through the model predictive control algorithm. Based on the historical trajectory information and environmental information of the trackless rubber-tyred vehicle, it is used to adjust the control parameters in a short time to deal with special situations, improve the control flexibility and adaptability of the trackless rubber-tyred vehicle, and realize the local path planning control of the trackless rubber-tyred vehicle; Step 5: Establish a collaborative control module, integrate the outputs of the long-term adaptive decision-making module and the local intelligent time-sensitive control module, and introduce a collaborative control weight parameter to dynamically allocate the control weight between the two; Step 6: The environmental perception module evaluates the environment surrounding the trackless rubber-tyred vehicle. Based on the dynamic weight optimization method of DDPG, the collaborative control weight parameters are intelligently and dynamically adjusted by adaptively learning the changes in the environment and vehicle status. Step 7: Based on the weight ratio of the long-term adaptive decision-making module and the local intelligent time-efficiency control module output by the collaborative control module in real time in step 6, while performing global path planning, local trajectory planning and trajectory tracking are performed based on the current local obstacle information, thereby adjusting and controlling the running trajectory of the rubber-tyred trackless vehicle to achieve smooth running trajectory.

4. The control method according to claim 3, characterized in that: The step 3 of establishing a long-term adaptive decision-making module for trackless rubber-tyred vehicles based on a genetic algorithm is specifically as follows: The genetic algorithm continuously optimizes the path through selection → crossover → mutation, and finally finds the optimal path for global path planning of trackless rubber-tyred vehicles. The specific process is as follows: Initialize the population to randomly generate N feasible paths, each path P is represented by a coordinate sequence: ， Calculate the comprehensive score for each path P: in ; Select Operation P: Crossover: Randomly select parents and , swap the fragments at position k; Mutation: For waypoints Random perturbations: ; New generation of population: final It terminates when it converges or reaches the maximum number of iterations, and generates the optimal path: .

5. The control method according to claim 3, characterized in that: The specific formula of the model predictive control algorithm in step 4 is: in is the running track state of the rubber-tyred trackless vehicle at the next moment, is the current state, Indicates the current speed and steering angle of the rubber-tyred trackless vehicle.

6. The control method according to claim 3, characterized in that: The collaborative control weight parameter in step 5 is α, which is used to calculate the weight. The specific formula is: Where: and are the outputs of the long-term adaptive decision-making module and the local intelligent time-sensitive control module respectively, α is the collaborative control weight parameter, and 0≤α≤1.

7. The control method according to claim 3, characterized in that: The DDPG dynamic weight optimization method in step 6 is specifically as follows: Initialize, select action Execute an action , the reward function is , the new state is ; Storage experience arrive ; Sampling small batches - ; Update Critic: Minimize Update Actors: Soft update TargetNetworks: Where s represents the current state of the rubber-tyred trackless vehicle, which is obtained by the surrounding environment information provided by the environment perception module; is the output action of the weight optimizer, which is the value of the collaborative control weight parameter α; r is the reward function, which represents the deviation between the actual trajectory and the planned trajectory and the smoothness of the trackless rubber-wheeled vehicle and the reverse trajectory during operation. The smaller the deviation and the smoother the running trajectory, the greater the reward; the weight optimizer selects the operation with the largest Q value according to the real-time environmental information fed back by the environmental perception module, and obtains the corresponding The value is the optimal value of the collaborative control weight parameter α.

Images