Mining robot path planning method based on artificial intelligence

Abstract

The invention discloses a mining robot path planning method based on artificial intelligence, relates to the technical field of artificial intelligence, and aims at solving the problem that the path of a mining robot can be planned by introducing an environment perception coefficient Rfx, an environment change coefficient Envc and a dynamic obstacle monitoring mechanism and combining real-time data processing of a dynamic obstacle total number Zzs, an obstacle distance Zjj, an obstacle relative speed Zsd and a robot steering radius Zbj. The problem that the traditional path planning cannot adapt to the sudden change of the mine environment is effectively solved; an obstacle avoidance path is generated by dynamically evaluating an environment change coefficient Envc, so that the safety and adaptability of path planning are improved; meanwhile, the emergency degree coefficient Urg of the assessment task is calculated, a path adjustment scheme is preferentially generated, and the emergency task response efficiency is improved; besides, whether a path optimization mode is started or not is judged through calculation and evaluation of a path energy consumption evaluation value Nex, it is guaranteed that tasks are completed smoothly, and the energy consumption risk is reduced; and the dynamic adaptability, the task allocation efficiency and the energy consumption management capability of the path planning of the mining robot are integrally improved.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and in particular to a mining robot path planning method based on artificial intelligence. Background Art

[0002] The background technology of the path planning method for mining robots can be traced back to the end of the last century. With the continuous advancement of mining automation, the complexity of the mining environment has prompted the gradual application of path planning technology in the field of mining robots. Early path planning mainly relied on preset routes, which was difficult to cope with the dynamically changing mining environment. With the improvement of sensor technology, artificial intelligence and computing power, the path planning method of mining robots has gradually developed characteristics based on environmental perception, autonomous decision-making and dynamic adjustment. In recent years, the introduction of deep learning and reinforcement learning algorithms has enabled mining robots to achieve efficient path optimization in complex mining environments, effectively avoid obstacles and improve operating efficiency.

[0003] However, the path planning method for mining robots has the following three technical disadvantages in the existing technology: 1. Insufficient dynamic adaptability of path planning: In traditional mining robot path planning methods, a common problem is that path planning cannot adapt to sudden changes in the mine environment in real time, such as dynamic obstacles, sudden changes in special terrain in the mine, etc. This will cause the robot to easily collide or the path cannot be executed, thereby reducing the efficiency and safety of task completion.

[0004] 2. Task priority is not enough to achieve optimal resource allocation: Existing path planning technologies usually lack dynamic analysis of the urgency of different robot tasks, and cannot adjust robot scheduling and path planning in time according to the urgency of the task. This will lead to waste of resources and delays in urgent tasks, and fail to fully utilize the overall resources of the mining area.

[0005] 3. Insufficient energy consumption assessment and power management: In traditional mining robot path planning methods, there is usually no detailed assessment of path energy consumption. Especially in complex mine environments, the robot may still try to complete the task when the power is insufficient, which increases the risk of mission failure and may also cause the robot to stay in a dangerous area and be unable to return. Summary of the invention

[0006] In view of the shortcomings of the prior art, the present invention provides a mining robot path planning method based on artificial intelligence, which solves the technical shortcomings of the background technology such as insufficient environmental perception, weak adaptability and limited intelligent decision-making ability.

[0007] To achieve the above objectives, the present invention is implemented through the following technical solutions: a mining robot path planning method based on artificial intelligence, comprising the following steps: S1. Pre-acquire a three-dimensional electronic map of the mining environment, combine it with the sensor data in the mine, and lock several potential obstacles and dynamically changing areas; then divide the mine area based on the three-dimensional electronic map to obtain several groups of mine sub-areas; then preset several groups of mining robots and mark them with serial numbers, and finally pair the mining robots with the mine sub-areas; S2. Real-time monitoring of the power state, position state and environmental perception information of the mining robot in the mine sub-area to generate a comprehensive state set, determine and evaluate the environmental perception coefficient Rfx of the mining robot at the current position based on the comprehensive state set, and screen out mining robots with insufficient environmental perception or plan a preliminary obstacle avoidance path for the mining robot according to the evaluation content of the environmental perception coefficient Rfx; S3. Based on the insufficient environmental perception of the mining robot obtained in step S2, combined with the sudden state changes of the working environment of the mine sub-area, the task processing data information of several groups of mining robots are monitored in real time, a task state set is constructed, and the current task urgency coefficient Urg of the mining robot with sequence number i is generated according to the task state set. The path planning priority is adjusted based on the task urgency coefficient Urg, and a dynamic path adjustment plan is generated for robots with urgent tasks first; S4. According to the path planning priority adjusted in step S3, dynamic obstacle monitoring is performed on the path of the robot with sequence number i, and the environment change coefficient Envc is generated and evaluated by combining the real-time environment perception data, dynamic obstacle data and the robot's motion model, and an advanced obstacle avoidance path is generated according to the evaluation content of the environment change coefficient Envc; S5. Simultaneously, the status data of the safety area and the working area in the mine sub-area are monitored and recorded. Combined with the current power consumption data of the robot with serial number i and the obstacle distribution data in the mine area, the path energy consumption evaluation value Nex for the robot to complete the current task is predicted and evaluated, and finally it is determined whether the mining robot can successfully complete the current task and whether it needs to enter the path optimization mode.

[0008] Preferably, S1 includes S11 and S12, specifically: S11, based on the environmental survey data of the mining area, extracting the geological data, mine structure data and geographic annotation information data of the mining area in advance, and then generating a three-dimensional electronic map of the mining area; the generation of the map is based on formatting the map; S12. After obtaining the three-dimensional electronic map, the system collects mine environment data in real time through multiple types of sensors arranged in the mine, including laser scanning, infrared sensors and ultrasonic sensors, and integrates the data with the three-dimensional electronic map; then, based on the information fed back by the sensors, it detects and locks potential obstacles, dynamically changing areas and special terrain in the mining area, and then marks the potential obstacles, dynamically changing areas and special terrain in the mining area, including the corresponding position coordinates and dynamic attributes.

[0009] Preferably, S2 includes S21 and S22, specifically: S21. Each robot obtains the current power, geographic location coordinates and surrounding environment conditions through embedded sensors; then extracts the remaining power value Eb, the obstacle density Obd at the current position, the mine environment light intensity Lum and the mine air dust concentration Dus in the comprehensive state set, and calculates the environmental perception coefficient Rfx of the current position of the mining robot through the following formula: .

[0010] Preferably, S22. By comparing and evaluating the preset environment perception threshold Q with the environment perception coefficient Rfx, robots with insufficient environment perception are screened out. The specific contents are as follows: When the environmental perception threshold Q is less than the environmental perception coefficient Rfx, it indicates that the robot's ability to perceive the current environment is abnormal and there is insufficient perception; the mining robot in the current mine sub-area is marked as abnormal and replaced or adjusted abnormally; When the environmental perception threshold Q ≥ environmental perception coefficient Rfx, it indicates that the robot has normal perception of the current environment; at this time, the current mining robot is planned to have a preliminary obstacle avoidance path, including dynamic obstacle avoidance based on the obstacle density Obd at the current position and the task target; Preferably, S3 includes S31 and S32, specifically: S31, the task status set includes the remaining task time Rsc, the task priority weight value Ryx, and the current task completion progress value Rjd; extract the task status set, and calculate the current task urgency coefficient Urg of the mining robot with sequence number i by the following formula: .

[0011] Preferably, S32, by comparing and evaluating the preset emergency threshold W with the task urgency coefficient Urg, the following evaluation content is generated: When the task urgency coefficient Urg ≥ the emergency threshold W, it indicates that the current task is urgent, and the robot will prioritize path adjustment. The system marks the robot as "task urgent" and immediately starts the dynamic path adjustment mode to generate a priority path for it to ensure that the task is completed in time.

[0012] When the task urgency coefficient Urg is less than the emergency threshold W, it indicates that the current task is not urgent, and the robot continues to execute according to the current path planning without additional path adjustment.

[0013] Preferably, S4 includes S41 and S42, specifically: S41, the real-time environment perception data, dynamic obstacle data and robot motion model data are aggregated, and after preprocessing and dimensionless processing, a dynamic obstacle data set is constructed; the total number of dynamic obstacles Zzs, obstacle distance Zjj, obstacle relative speed Zsd and robot turning radius Zbj in the dynamic obstacle data set are extracted, and the environment change coefficient Envc is generated by the following formula:

[0014] In the formula, represents the distance between the mining robot and the jth dynamic obstacle, represents the relative speed between the robot and the jth dynamic obstacle, Represents the turning radius of the mining robot when interacting with the jth obstacle.

[0015] Preferably, S42, compare and evaluate the preset safety threshold Es with the environmental variation coefficient Envc, and the specific evaluation contents are as follows: If the environmental change coefficient Envc> the preset safety threshold Es, the system determines that the robot's current path is risky and cannot pass safely; at this time, the system automatically generates an advanced obstacle avoidance path for the robot, including avoiding high-risk areas first and choosing a safe path with a lower obstacle density; If the environmental change coefficient Envc ≤ the preset safety threshold Es, the system determines that there is no risk in the current path of the robot, and there is no need to generate a new path.

[0016] Preferably, S5 includes S51 and S52, specifically: S51, wherein the status data of the safety area and the operating area in the mine sub-area include the passability rate Pav and the regional height difference Alt, and the current power consumption related data and the obstacle distribution related data in the mine area include the power consumption per meter distance Edi and the obstacle height Obh; after extracting the passability rate Pav, the regional height difference Alt, the power consumption per meter distance Edi and the obstacle height Obh for dimensionless processing, the path energy consumption evaluation value Nex is calculated by the following formula: .

[0017] Preferably, S52: Evaluate by using the preset path energy consumption threshold R and the path energy consumption evaluation value Nex. The specific contents are as follows: When the path energy consumption evaluation value Nex≤path energy consumption threshold R, it indicates that the energy consumption of the robot under the current path planning meets the threshold requirement; at this time, the system determines that the robot completes the task within the power range and continues to perform the task along the current path without path adjustment; When the path energy consumption evaluation value Nex> the path energy consumption threshold R, it indicates that the energy consumption of the robot under the current path planning cannot meet the threshold requirement; at this time, the system determines that the robot's power is insufficient to support the task completion; the system marks the robot as "path optimization requirement" and starts the path optimization mode.

[0018] The present invention provides a mining robot path planning method based on artificial intelligence. It has the following beneficial effects: (1) This artificial intelligence-based mining robot path planning method effectively solves the problem of insufficient dynamic adaptability of path planning by introducing the environmental perception coefficient Rfx, the environmental change coefficient Envc and the dynamic obstacle monitoring mechanism. The system monitors the potential dynamic change areas and special terrains in the mine sub-areas in real time, and generates the environmental change coefficient Envc after dimensionless processing of the total number of dynamic obstacles Zzs, obstacle distance Zjj, obstacle relative speed Zsd and robot turning radius Zbj based on sensor feedback data. According to the comparative evaluation results of Envc and the safety threshold Es, an obstacle avoidance path is dynamically generated to avoid collision or task interruption due to the robot's inability to recognize environmental changes. This improvement improves the path planning accuracy and safety of mining robots in complex mine environments.

[0019] (2) This is an artificial intelligence-based mining robot path planning method. To address the problem that task priority is not sufficient to achieve optimal resource allocation, this technical solution proposes a dynamic adjustment mechanism for the task urgency coefficient Urg; by extracting the task remaining time Rsc, task priority weight value Ryx, and current task completion progress value Rjd from the task status set and performing formula calculations; combining the task urgency threshold W to dynamically evaluate the task urgency coefficient Urg, when the task urgency coefficient Urg is greater than or equal to the task urgency threshold W, the robot is marked as "task urgent" and a dynamic path adjustment plan is generated preferentially; this method realizes the refined management of mining robot task allocation, significantly improves the response speed of emergency tasks, and optimizes the overall utilization efficiency of mining resources.

[0020] (3) This is a path planning method for mining robots based on artificial intelligence. In order to solve the problem of insufficient energy consumption evaluation and power management, this scheme proposes a dynamic prediction method for the path energy consumption evaluation value Nex; the path energy consumption evaluation value Nex is calculated by dimensionlessly processing the energy consumption per meter distance Edi, obstacle height Obh, passability Pav and regional height difference Alt; the path energy consumption threshold R is combined for evaluation. When the path energy consumption evaluation value Nex is greater than the path energy consumption threshold R, the system marks the robot as "path optimization requirement" and starts the path optimization mode to avoid task failure or the robot staying in a dangerous area due to insufficient power; this method ensures the task completion rate and high efficiency of energy consumption management of mining robots, while reducing the safety hazards of mine operations. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 A schematic diagram of the steps of a mining robot path planning method based on artificial intelligence of the present invention; Figure 2 The present invention is a schematic diagram of the scheme code of a mining robot path planning method based on artificial intelligence. DETAILED DESCRIPTION

[0022] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0023] Example 1 See also Figure 1 The present invention provides a mining robot path planning method based on artificial intelligence, comprising the following steps: S1. Pre-acquire a three-dimensional electronic map of the mining environment, combine it with the sensor data in the mine, and lock several potential obstacles and dynamically changing areas; then divide the mine area based on the three-dimensional electronic map to obtain several groups of mine sub-areas; then preset several groups of mining robots and mark them with serial numbers, and finally pair the mining robots with the mine sub-areas; S2. Real-time monitoring of the power state, position state and environmental perception information of the mining robot in the mine sub-area to generate a comprehensive state set, determine and evaluate the environmental perception coefficient Rfx of the mining robot at the current position based on the comprehensive state set, and screen out mining robots with insufficient environmental perception or plan a preliminary obstacle avoidance path for the mining robot according to the evaluation content of the environmental perception coefficient Rfx; S3. Based on the insufficient environmental perception of the mining robot obtained in step S2, combined with the sudden state changes of the working environment of the mine sub-area, the task processing data information of several groups of mining robots are monitored in real time, a task state set is constructed, and the current task urgency coefficient Urg of the mining robot with sequence number i is generated according to the task state set. The path planning priority is adjusted based on the task urgency coefficient Urg, and a dynamic path adjustment plan is generated for robots with urgent tasks first; S4. According to the path planning priority adjusted in step S3, dynamic obstacle monitoring is performed on the path of the robot with sequence number i, and the environment change coefficient Envc is generated and evaluated by combining the real-time environment perception data, dynamic obstacle data and the robot's motion model, and an advanced obstacle avoidance path is generated according to the evaluation content of the environment change coefficient Envc; S5. Simultaneously, the status data of the safety area and the working area in the mine sub-area are monitored and recorded. Combined with the current power consumption data of the robot with serial number i and the obstacle distribution data in the mine area, the path energy consumption evaluation value Nex for the robot to complete the current task is predicted and evaluated, and finally it is determined whether the mining robot can successfully complete the current task and whether it needs to enter the path optimization mode.

[0024] In this embodiment, step S1 acquires a three-dimensional electronic map of the mining environment and combines it with sensor data in the mine to lock potential obstacles and dynamically changing areas, and divides the mine into sub-areas and pairs them with mining robots, providing basic regional division and preliminary environmental recognition capabilities for path planning; step S2 generates a comprehensive state set by real-time monitoring of the power state, position state and environmental perception information of the mining robot, and calculates the environmental perception coefficient Rfx of the current position, screens out robots with insufficient environmental perception and generates a preliminary obstacle avoidance path for them, thereby improving the perception accuracy and safety of the robot in a complex mining environment; step S3 constructs a task state set and generates a task urgency coefficient Urg, and adjusts the path planning priority based on the task urgency At the level, a dynamic path adjustment plan is generated for robots with urgent tasks, so that they can respond to sudden changes in the mining area in a timely manner and improve the efficiency and adaptability of task completion; step S4 generates an environmental change coefficient Envc according to dynamic obstacle data, real-time environmental perception data and robot motion model, and generates an advanced obstacle avoidance path through evaluation, which effectively reduces the collision risk of the robot in a dynamic environment and ensures the safety of the path; step S5 monitors and records the status data of the safety area and the working area in the mine sub-area, combines the current power consumption data and obstacle distribution data, and calculates the path energy consumption evaluation value Nex, so as to evaluate whether the robot can successfully complete the task or needs to enter the path optimization mode, so as to ensure the efficient completion of the task and the optimal choice of the path under power constraints.

[0025] Example 2 S1 includes S11 and S12, specifically: S11, based on the environmental survey data of the mining area, extracting the geological data, mine structure data and geographic annotation information data of the mining area in advance, and then generating a three-dimensional electronic map of the mining area; the generation of the map is based on formatting the map; S12. After obtaining the three-dimensional electronic map, the system collects mine environment data in real time through multiple types of sensors arranged in the mine, including laser scanning, infrared sensors and ultrasonic sensors, and integrates the data with the three-dimensional electronic map; then, based on the information fed back by the sensors, it detects and locks potential obstacles, dynamically changing areas and special terrain in the mining area, and then marks the potential obstacles, dynamically changing areas and special terrain in the mining area, including the corresponding position coordinates and dynamic attributes.

[0026] In this embodiment, step S1 provides comprehensive environmental data support and dynamic monitoring basis for the mining robot path planning by including the operations of S11 and S12; in S11, by extracting the geological data, mine structure data and geographic annotation information of the mining area, a formatted three-dimensional electronic map is generated, so that the system has the basic geographic structure information of the mining area; In S12, the system uses laser scanning, infrared sensors and ultrasonic sensors to collect mine environment data in real time, and integrates it with the three-dimensional electronic map. It combines sensor feedback information to detect and lock potential obstacles, dynamically changing areas and special terrain in the mine area, and marks the location coordinates and dynamic attributes of these key areas, so that the system can continuously obtain accurate and dynamic spatial information in the complex mining environment. This step plays a key role in the path planning method, enabling the robot to predict and avoid obstacles, adapt to environmental changes in real time, and improve the accuracy of path planning and the safety of task execution. The specific process of detecting and locking potential obstacles, dynamically changing areas and special terrain in the mining area is as follows: the laser scanning sensor scans the three-dimensional space of the mining area, generates preliminary contour data of the obstacles, and measures their position coordinates; then, the infrared sensor captures the changes in thermal signals to identify the movement trajectory of the heat source in the dynamic area; subsequently, the ultrasonic sensor detects the surface concave and convex features in the mine through sound wave reflection to identify the special terrain; all sensor data are transmitted to the central processing unit in real time, and multi-source information is integrated using data fusion technology, and the movement trend of obstacles, changes in heat source intensity and terrain complexity are analyzed through a specific dynamic change model; the system classifies and marks the analysis results, including static obstacles, boundary attributes of dynamically changing areas and risk levels of special terrain, and stores their locations and dynamic attributes in a three-dimensional electronic map to ensure that the analysis results are called in real time during the robot path planning process, thereby improving the perception accuracy and safety of the mine environment.

[0027] Example 3 S2 includes S21 and S22, specifically: S21. Each robot obtains the current power, geographic location coordinates and surrounding environment conditions through embedded sensors; then extracts the remaining power value Eb, the obstacle density Obd at the current position, the mine environment light intensity Lum and the mine air dust concentration Dus in the comprehensive state set, and calculates the environmental perception coefficient Rfx of the current position of the mining robot through the following formula:

[0028] S22. By comparing and evaluating the preset environment perception threshold Q with the environment perception coefficient Rfx, robots with insufficient environment perception are screened out. The specific contents are as follows: When the environmental perception threshold Q is less than the environmental perception coefficient Rfx, it indicates that the robot's ability to perceive the current environment is abnormal and there is insufficient perception; the mining robot in the current mine sub-area is marked as abnormal and replaced or adjusted abnormally; When the environmental perception threshold Q ≥ environmental perception coefficient Rfx, it indicates that the robot has normal perception of the current environment; at this time, the current mining robot is planned to have a preliminary obstacle avoidance path, including dynamic obstacle avoidance based on the obstacle density Obd at the current position and the task target; S3 includes S31 and S32, specifically: S31, the task status set includes the remaining task time Rsc, the task priority weight value Ryx, and the current task completion progress value Rjd; extract the task status set, and calculate the current task urgency coefficient Urg of the mining robot with sequence number i by the following formula:

[0029] S32, by comparing and evaluating the preset emergency threshold W with the task urgency coefficient Urg, the following evaluation content is generated: When the task urgency coefficient Urg ≥ the emergency threshold W, it indicates that the current task is urgent, and the robot will prioritize path adjustment. The system marks the robot as "task urgent" and immediately starts the dynamic path adjustment mode to generate a priority path for it to ensure that the task is completed in time; When the task urgency coefficient Urg is less than the emergency threshold W, it indicates that the current task is not urgent, and the robot continues to execute according to the current path planning without additional path adjustment.

[0030] S4 includes S41 and S42, specifically: S41, the real-time environment perception data, dynamic obstacle data and robot motion model data are aggregated, and after preprocessing and dimensionless processing, a dynamic obstacle data set is constructed; the total number of dynamic obstacles Zzs, obstacle distance Zjj, obstacle relative speed Zsd and robot turning radius Zbj in the dynamic obstacle data set are extracted, and the environment change coefficient Envc is generated by the following formula:

[0031] In the formula, represents the distance between the mining robot and the jth dynamic obstacle, represents the relative speed between the robot and the jth dynamic obstacle, Represents the turning radius of the mining robot when interacting with the jth obstacle.

[0032] S42, compare and evaluate the preset safety threshold Es with the environmental variation coefficient Envc, and the specific evaluation contents are as follows: If the environmental change coefficient Envc> the preset safety threshold Es, the system determines that the robot's current path is risky and cannot pass safely; at this time, the system automatically generates an advanced obstacle avoidance path for the robot, including avoiding high-risk areas first and choosing a safe path with a lower obstacle density; If the environmental change coefficient Envc ≤ the preset safety threshold Es, the system determines that there is no risk in the current path of the robot, and there is no need to generate a new path.

[0033] In this embodiment, accurate perception, adaptive adjustment and intelligent path optimization of the robot in the mine area are achieved through various steps; in S2, each robot obtains the current power, geographical location coordinates and surrounding environmental conditions through embedded sensors, and extracts the remaining power value Eb, the obstacle density Obd at the current position, the mine environment light intensity Lum and the mine air dust concentration Dus in the comprehensive state set, and generates the environmental perception coefficient Rfx through calculation. This coefficient is used to evaluate the current perception ability of the robot to ensure timely replacement or adjustment of the path when the perception is insufficient; wherein, the obstacle density Obd represents a single The distribution density of the number of obstacles in the location area is used to reflect the complexity of the mine environment. The mine environment data is obtained in real time through multiple types of sensors deployed in the mine, including laser scanners, ultrasonic sensors and infrared sensors. The number of obstacles in the sensor coverage area is counted, and the number of obstacles per unit area is calculated in combination with the geographic location coordinates. The environmental perception threshold Q is obtained through experimental calibration, usually set according to the complexity of the mine environment, and the value range is generally between 0.1 and 1. The specific value of the environmental perception coefficient Rfx is calculated based on the parameters, and the general value range is 0.0~1.0.

[0034] S3 uses the remaining task time Rsc, the task priority weight value Ryx and the current task completion progress value Rjd to generate the task urgency coefficient Urg, and compares it with the preset emergency threshold W to determine whether priority path adjustment is needed to ensure that the emergency task can be completed in time; the task urgency coefficient Urg is calculated by the remaining task time Rsc (in seconds) and the task priority weight value Ryx (0-1), usually in the range of 0 to 10; the emergency threshold W is set according to the time sensitivity of the mining area task, usually in the range of 5 to 8; at the same time, the dynamic path adjustment mode includes combining the robot's motion model data and using The dynamic path planning algorithm generates a new path set in real time; then, by evaluating the safety, passability and energy consumption of each path in the path set, the path with low obstacle density, minimum risk in the dynamically changing area, and energy consumption evaluation value Nex less than or equal to the path energy consumption threshold R is selected as the priority path; at the same time, the system marks the high-risk areas in real time and notifies the robot to bypass these areas; during the path execution process, the system continuously monitors the changes in dynamic obstacles and the status of the robot. If new environmental changes occur, the path parameters are dynamically adjusted to ensure that the robot is always in the optimal path state; the following is a specific example after the dynamic path adjustment mode is started: Case 1: The robot's original path was planned to pass through a mine tunnel with a large slope, a high density of obstacles, and a landslide risk in this area. After the system detected that this path was not feasible, it chose to bypass a safer channel. Even though this path had a higher energy consumption, it still gave priority to this path to ensure safety and complete the task in a timely manner; Case 2: The system detects a new obstacle (such as fallen rocks) on another path in the mine. After the path is replanned, the robot is guided to bypass the obstacle. The updated path can avoid the obstacle and ensure the timeliness of the task, completing the shortest path planning; In S4, the real-time environment perception data, dynamic obstacle data, and robot motion model data are aggregated, and a dynamic obstacle data set is constructed after preprocessing and dimensionless processing. The specific construction process of the dynamic obstacle data set is as follows: Using multimodal data fusion technology, the data collected by laser scanning, ultrasonic sensors and infrared sensors are preprocessed, including data cleaning, noise reduction and format unification; the dimensionless method is used to process the distance, speed and number of dynamic obstacles in a dimensionless manner so as to uniformly construct a dynamic obstacle data set; the specific dimensionless processing is as follows: , X represents the original parameter value, Xmin represents the minimum value of the original parameter, and Xmax represents the maximum value of the original parameter.

[0035] The environmental change coefficient Envc is generated by the total number of dynamic obstacles Zzs, obstacle distance Zjj, obstacle relative speed Zsd and robot turning radius Zbj, so as to evaluate the safety of the path and generate an advanced obstacle avoidance path when necessary, to ensure the safety and path continuity of the robot in a dynamic and complex environment; by collecting and evaluating each subordinate parameter, a comprehensive monitoring of the mining operating environment is achieved, and accurate data support in path planning is provided, so that the system has adaptability and intelligent decision-making capabilities, ensuring efficient completion of robot tasks and safe operation; among them, based on the total number of obstacles Zzs (in number), obstacle distance Zjj (in meter, range 0-5 meters) and robot turning radius Zbj (in meter, range 0.1-10 meters) in the dynamic obstacle data set; the safety threshold Es is preset according to the safety level and risk tolerance of the mine environment, and usually takes a value range of 5~7.

[0036] Example 4 S5 includes S51 and S52, specifically: S51, wherein the status data of the safety area and the operating area in the mine sub-area include the passability rate Pav and the regional height difference Alt, and the current power consumption related data and the obstacle distribution related data in the mine area include the power consumption per meter distance Edi and the obstacle height Obh; after extracting the passability rate Pav, the regional height difference Alt, the power consumption per meter distance Edi and the obstacle height Obh for dimensionless processing, the path energy consumption evaluation value Nex is calculated by the following formula:

[0037] The evaluation is performed by presetting the path energy consumption threshold R and the path energy consumption evaluation value Nex. The specific contents are as follows: When the path energy consumption evaluation value Nex≤path energy consumption threshold R, it indicates that the energy consumption of the robot under the current path planning meets the threshold requirement; at this time, the system determines that the robot completes the task within the power range and continues to perform the task along the current path without path adjustment; When the path energy consumption evaluation value Nex> the path energy consumption threshold R, it indicates that the energy consumption of the robot under the current path planning cannot meet the threshold requirement; at this time, the system determines that the robot's power is insufficient to support the task completion; the system marks the robot as "path optimization requirement" and starts the path optimization mode.

[0038] In this embodiment, S5 is used to realize energy consumption optimization and path decision of the robot in the mining area to ensure efficient completion of the task; in S51, the system collects the status data of the safety area and the working area in the mine sub-area, wherein the passability Pav and the area height difference Alt provide the passage difficulty and slope information of the robot under different path conditions, the electric energy consumption per meter distance Edi represents the unit energy consumption of the robot under standard path conditions, and the obstacle height Obh reflects the potential energy consumption impact of obstacles in the path on the robot; by dimensionlessly processing these parameters, the system calculates the path energy consumption evaluation value Nex, thereby quantifying the total energy consumption required for the path in actual execution; Among them, the value range of the path energy consumption threshold R is mainly determined by the complexity of the mine environment, the battery capacity of the robot and the task completion requirements, and is usually set according to the maximum energy consumption limit of the task; the specific values ​​are generally: the energy consumption unit is usually J / m joule per meter. In the mine environment, the path energy consumption threshold R typically ranges from 10 to 100 J / m, and the value is higher in complex environments. The setting method is to evaluate the average energy consumption of the unit path of the mining robot in different environments through experiments or simulations, and then add a certain safety margin to obtain it; the path energy consumption evaluation value Nex ranges from 15 to 120 J / m, and the specific value varies depending on the path length and environmental complexity; In S52, the system compares and evaluates the path energy consumption evaluation value Nex with the preset path energy consumption threshold R to determine whether the current path is feasible within the robot's power range. After determining that the robot has insufficient power, the system marks it as "path optimization requirement" and starts the path optimization mode to generate a low-energy consumption path for the robot. This method enables the mining robot to complete tasks in an environment with limited power through accurate energy consumption evaluation, thereby improving the energy efficiency of path planning and the reliability of task completion. Among them, the passage rate Pav represents the probability that a robot can pass safely in a certain area within a unit time. The specific quantification method is: within a specific time window, count the number of obstacles in the area, including static and dynamic obstacles, and the movement speed Zsd of dynamic obstacles; then combine the environmental perception data to calculate the proportion of the area covered by obstacles to the total area; take the obstacle density Obd and the movement frequency of dynamic obstacles as the core parameters, and use the following formula to calculate: , where Aobs represents the total area covered by obstacles, and Atotal represents the total area of ​​the region; the value is in [0, 1], where the closer the value is to 1, the higher the pass rate; The height difference Alt represents the vertical height change between the starting point and the end point in the robot path planning; the specific acquisition method is as follows: Use laser scanning equipment or high-precision terrain mapping equipment deployed in the mine to obtain three-dimensional terrain data; extract the height values ​​z1 and z2 in the starting coordinates (x1, y1, z1) and the end coordinates (x2, y2, z2) of the robot's current path; calculate the height difference using the following formula: ,Range: The value range of the height difference Alt is usually between 0 and 10 meters according to the terrain characteristics of the mining area, and the specific range is determined by the complexity of the mine environment.

[0039] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A mining robot path planning method based on artificial intelligence, characterized in that: The following steps are involved: S1. Pre-acquire a three-dimensional electronic map of the mining environment, combine it with the sensor data in the mine, and lock several potential obstacles and dynamically changing areas; then divide the mine area based on the three-dimensional electronic map to obtain several groups of mine sub-areas; then preset several groups of mining robots and mark them with serial numbers, and finally pair the mining robots with the mine sub-areas; S2. Real-time monitoring of the power state, position state and environmental perception information of the mining robot in the mine sub-area to generate a comprehensive state set, determine and evaluate the environmental perception coefficient Rfx of the mining robot at the current position based on the comprehensive state set, and screen out mining robots with insufficient environmental perception or plan a preliminary obstacle avoidance path for the mining robot according to the evaluation content of the environmental perception coefficient Rfx; S3. Based on the insufficient environmental perception of the mining robot obtained in step S2, combined with the sudden state changes of the working environment of the mine sub-area, the task processing data information of several groups of mining robots are monitored in real time, a task state set is constructed, and the current task urgency coefficient Urg of the mining robot with sequence number i is generated according to the task state set. The path planning priority is adjusted based on the task urgency coefficient Urg, and a dynamic path adjustment plan is generated for robots with urgent tasks first; S4. According to the path planning priority adjusted in step S3, dynamic obstacle monitoring is performed on the path of the robot with sequence number i, and the environment change coefficient Envc is generated and evaluated by combining the real-time environment perception data, dynamic obstacle data and the robot's motion model, and an advanced obstacle avoidance path is generated according to the evaluation content of the environment change coefficient Envc; S5. Simultaneously, the status data of the safety area and the working area in the mine sub-area are monitored and recorded. Combined with the current power consumption data of the robot with serial number i and the obstacle distribution data in the mine area, the path energy consumption evaluation value Nex for the robot to complete the current task is predicted and evaluated, and finally it is determined whether the mining robot can successfully complete the current task and whether it needs to enter the path optimization mode.

2. The path planning method for a mining robot based on artificial intelligence according to claim 1, characterized in that: S1 includes S11 and S12, specifically: S11, based on the environmental survey data of the mining area, extracting the geological data, mine structure data and geographic annotation information data of the mining area in advance, and generating a three-dimensional electronic map of the mining area; S12. After obtaining the three-dimensional electronic map, the system collects mine environment data in real time through multiple types of sensors arranged in the mine, including laser scanning, infrared sensors and ultrasonic sensors, and integrates the data with the three-dimensional electronic map; based on the information fed back by the sensors, the system detects and locks potential obstacles, dynamically changing areas and special terrain in the mining area, and then marks the potential obstacles, dynamically changing areas and special terrain in the mining area, including the corresponding position coordinates and dynamic attributes.

3. The path planning method for a mining robot based on artificial intelligence according to claim 1, characterized in that: S2 includes S21 and S22, specifically: S21. Each robot obtains the current power, geographic location coordinates and surrounding environment conditions through embedded sensors; then extracts the remaining power value Eb, the obstacle density Obd at the current position, the mine environment light intensity Lum and the mine air dust concentration Dus in the comprehensive state set, and calculates the environmental perception coefficient Rfx of the current position of the mining robot through the following formula: 。 4. The path planning method for a mining robot based on artificial intelligence according to claim 1, characterized in that: S22. By comparing and evaluating the preset environment perception threshold Q with the environment perception coefficient Rfx, robots with insufficient environment perception are screened out. The specific contents are as follows: When the environmental perception threshold Q is less than the environmental perception coefficient Rfx, it indicates that the robot's ability to perceive the current environment is abnormal and there is insufficient perception; the mining robot in the current mine sub-area is marked as abnormal and replaced or adjusted abnormally; When the environmental perception threshold Q ≥ environmental perception coefficient Rfx, it indicates that the robot's ability to perceive the current environment is normal; at this time, the current mining robot is planned to have a preliminary obstacle avoidance path, including dynamic obstacle avoidance based on the obstacle density Obd at the current position and the task target.

5. The path planning method for a mining robot based on artificial intelligence according to claim 1, characterized in that: S3 includes S31 and S32, specifically: S31, the task status set includes the remaining task time Rsc, the task priority weight value Ryx, and the current task completion progress value Rjd; extract the task status set, and calculate the current task urgency coefficient Urg of the mining robot with sequence number i by the following formula: 。 6. The path planning method for a mining robot based on artificial intelligence according to claim 1, characterized in that: S32, by comparing and evaluating the preset emergency threshold W with the task urgency coefficient Urg, the following evaluation content is generated: When the task urgency coefficient Urg ≥ the emergency threshold W, it indicates that the current task is urgent. The robot will prioritize path adjustment at this time. The system will mark the robot as "task urgent" and immediately start the dynamic path adjustment mode to generate a priority path for it. When the task urgency coefficient Urg is less than the emergency threshold W, it indicates that the current task is not urgent, and the robot continues to execute according to the current path planning without additional path adjustment.

7. The path planning method for a mining robot based on artificial intelligence according to claim 1, characterized in that: S4 includes S41 and S42, specifically: S41, summarizing the real-time environment perception data, dynamic obstacle data, and robot motion model data, and constructing a dynamic obstacle data set after preprocessing and dimensionless processing; The total number of dynamic obstacles Zzs, obstacle distance Zjj, obstacle relative speed Zsd and robot turning radius Zbj in the dynamic obstacle data set are extracted, and the environment change coefficient Envc is generated by the following formula: In the formula, represents the distance between the mining robot and the jth dynamic obstacle, represents the relative speed between the robot and the jth dynamic obstacle, Represents the turning radius of the mining robot when interacting with the jth obstacle.

8. The path planning method for a mining robot based on artificial intelligence according to claim 1, characterized in that: S42, compare and evaluate the preset safety threshold Es with the environmental variation coefficient Envc, and the specific evaluation contents are as follows: If the environmental change coefficient Envc> the preset safety threshold Es, the system determines that the robot's current path is risky and cannot pass safely; at this time, the system automatically generates an advanced obstacle avoidance path for the robot, including avoiding high-risk areas first and choosing a safe path with a lower obstacle density; If the environmental change coefficient Envc ≤ the preset safety threshold Es, the system determines that there is no risk in the current path of the robot, and there is no need to generate a new path.

9. The path planning method for a mining robot based on artificial intelligence according to claim 1, characterized in that: S5 includes S51 and S52, specifically: S51, wherein the status data of the safety area and the operating area in the mine sub-area include the passability rate Pav and the regional height difference Alt, and the current power consumption related data and the obstacle distribution related data in the mine area include the power consumption per meter distance Edi and the obstacle height Obh; after extracting the passability rate Pav, the regional height difference Alt, the power consumption per meter distance Edi and the obstacle height Obh for dimensionless processing, the path energy consumption evaluation value Nex is calculated by the following formula: 。 10. The path planning method for a mining robot based on artificial intelligence according to claim 1, characterized in that: S52: Evaluate by using the preset path energy consumption threshold R and the path energy consumption evaluation value Nex. The specific contents are as follows: When the path energy consumption evaluation value Nex≤path energy consumption threshold R, it indicates that the energy consumption of the robot under the current path planning meets the threshold requirement; at this time, the system determines that the robot completes the task within the power range and continues to perform the task along the current path without path adjustment; When the path energy consumption evaluation value Nex> the path energy consumption threshold R, it indicates that the energy consumption of the robot under the current path planning cannot meet the threshold requirement; at this time, the system determines that the robot's power is insufficient to support the task completion; the system marks the robot as "path optimization requirement" and starts the path optimization mode.

Images