Unmanned vehicle cluster collaborative scheduling system, method, device and medium

Abstract

The invention relates to an unmanned vehicle group cooperative scheduling system, method, device and medium, and relates to the technical field of automatic guided vehicle scheduling, the system comprises a central scheduling platform and a plurality of vehicle-mounted intelligent units corresponding to AGVs one by one, the central scheduling platform obtains a pebble coal carrying task request and AGV state information, and the AGV state information is sent to the central scheduling platform; determining at least one target AGV based on the task request, the state information of each AGV and a preset AGV capability evaluation model, and performing global path planning on each target AGV and sending path information; and the vehicle-mounted intelligent unit acquires local environment data in real time, performs local optimization on the global path to generate an obstacle avoidance track, and controls the target AGV to run according to the local obstacle avoidance track. And the central scheduling platform also performs conflict detection and resolution and processes an arbitration request. According to the method, the AGV cluster can be reasonably scheduled, the task execution efficiency is improved, the AGV conflict is effectively processed, and safe and efficient operation of the AGV is guaranteed.

Description

Technical Field

[0001] This invention relates to the field of automated guided vehicle (AGV) scheduling technology, and in particular to an unmanned vehicle cluster collaborative scheduling system, method, equipment, and medium. Background Technology

[0002] With the continuous development of industrial automation control technology, AGV (Automated Guided Vehicle) technology has demonstrated enormous potential in industrial logistics automation scenarios. In thermal power plants, the handling of coal dust discharged from coal mills is a critical link in ensuring the stable operation of the units, and efficient and safe handling methods are essential for the stable production of the power plant. With the advancement of industrial intelligence, the demand for intelligent scheduling and management, such as multi-vehicle collaborative operation, dynamic path planning, conflict resolution, and task optimization allocation, is increasing. AGV cluster collaborative scheduling systems suitable for complex industrial environments have become a research and application hotspot.

[0003] In traditional coal and stone handling at thermal power plants, reliance is primarily placed on manual labor or fixed mechanical equipment. Manual labor requires a significant workforce, with workers frequently moving back and forth within the power plant environment, resulting in high labor intensity. While fixed mechanical equipment can alleviate the burden on manpower to some extent, it lacks flexibility and can only operate according to pre-set methods. In recent years, AGV technology has been applied in industrial logistics. Existing AGV scheduling systems mostly employ fixed path planning or simple traffic rule methods. Fixed path planning guides AGVs along fixed trajectories based on pre-set maps and routes; simple traffic rule methods establish basic driving rules for AGVs, such as yielding rules, to avoid conflicts between vehicles.

[0004] However, in the complex environment of thermal power plants, when multiple AGVs operate in confined spaces, fixed-path planning or simple traffic rule methods are prone to path conflicts and traffic congestion. Because the work area is an open environment, with other plant personnel and vehicles also operating in the same space, the real-time requirements for task scheduling are higher. Existing methods lack dynamic adjustment capabilities and struggle to adapt to the real-time changes in workload at thermal power plants. Simultaneously, the emissions of coking coal from various coal mills in thermal power plants are severely unbalanced and fluctuate greatly, making them difficult to predict; existing methods cannot effectively address this situation. Furthermore, the high temperature, high humidity, and high dust conditions within power plants affect sensor accuracy and equipment stability; existing systems have poor environmental adaptability, and compatibility with existing DCS, SIS, and other systems in power plants also presents challenges. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to provide an unmanned vehicle cluster collaborative scheduling system, method, device and medium, aiming to solve at least one of the above-mentioned technical problems.

[0006] The technical solution of the present invention to solve the above-mentioned technical problems is as follows: Firstly, this application provides an unmanned vehicle cluster collaborative scheduling system, which adopts the following technical solution: An unmanned vehicle cluster collaborative scheduling system includes a central scheduling platform and multiple on-board intelligent units corresponding one-to-one with AGVs; The central scheduling platform is used to obtain task requests for transporting stones and coal, as well as the status information of each AGV in the AGV cluster. The task request includes the task location and task priority, and the status information includes the location, battery level, average speed, and current number of tasks. The central scheduling platform is used to determine at least one target AGV based on the task request, the status information of each AGV, and a preset AGV capability evaluation model. The central scheduling platform is used to perform global path planning for each target AGV based on the task request, the status information of each target AGV and the preset path planning algorithm, to obtain global reference path information for each target AGV, and to send each global reference path information to the vehicle-mounted intelligent unit corresponding to each target AGV. The on-board intelligent unit corresponding to any of the target AGVs is used to acquire local environmental data of the target AGV in real time during the process of the target AGV driving based on the global reference path information, perform local real-time optimization of the global reference path information based on the local environmental data, generate a local obstacle avoidance trajectory, and control the target AGV to drive according to the local obstacle avoidance trajectory.

[0007] The beneficial effects of this invention are as follows: the central scheduling platform determines the target AGV based on the task request and AGV status information, combined with the capability assessment model, and performs global path planning, providing global reference path information for the target AGV, thereby realizing intelligent task allocation and path planning for multi-AGV collaborative operations; the onboard intelligent unit can perform local real-time optimization of the global reference path information based on local environmental data, generate local obstacle avoidance trajectories, improve the system's adaptability and safety in complex environments, and enhance the overall system efficiency and resource utilization.

[0008] Based on the above technical solution, the present invention can be further improved as follows.

[0009] Furthermore, the central scheduling platform is also used for: Based on the global reference path information and status information of each AGV, conflict detection is performed on each AGV to obtain the conflict detection result. Based on the conflict detection results and a pre-set traffic rule base, a conflict resolution strategy is generated for one or more AGVs among all the AGVs involved in the conflict detection results. The strategy includes at least one of speed adjustment, path replanning, or waiting at a safe point. The conflict resolution strategy is sent to the onboard intelligent unit of the AGV corresponding to the conflict resolution strategy.

[0010] The beneficial effects of adopting the above-mentioned further solution are as follows: The central scheduling platform can perform conflict detection based on the global reference path information and status information of AGVs, obtain the conflict detection results, and then generate conflict resolution strategies in combination with the pre-set traffic rule base, including speed adjustment, path replanning, or waiting at a safe point. The strategies are sent to the on-board intelligent unit of the corresponding AGV, thereby solving the problem of multi-AGV path conflict and traffic congestion, improving the overall system efficiency and resource utilization, and ensuring the efficient, safe and smooth operation of the AGV cluster in complex environments.

[0011] Furthermore, the on-board intelligent unit is also used to determine, based on the task request, whether the execution of the conflict resolution strategy by the AGV will affect the time for the AGV to execute the current task. If the execution of the conflict resolution strategy by the AGV will affect the time for the AGV to execute the current task, then an arbitration request is sent to the central scheduling platform. The central scheduling platform is also used to make arbitration decisions based on the arbitration request and the task priority corresponding to the AGV.

[0012] The beneficial effects of adopting the above-mentioned further solution are: the on-board intelligent unit determines whether the conflict resolution strategy executed by the AGV affects the current task time. If it does, it sends an arbitration request to the central scheduling platform. The central scheduling platform makes an arbitration decision based on the arbitration request and task priority, which can ensure that the task is completed on time while resolving AGV conflicts, improve the rationality of system scheduling and the reliability of task execution, and avoid the impact of conflict resolution on task execution efficiency.

[0013] Furthermore, the central scheduling platform is used to perform conflict detection on each AGV based on the global reference path information and status information of each AGV. When obtaining the conflict detection result, it is specifically used for: Based on the global reference path information of each AGV in the AGV cluster, each global reference path information is decomposed into multiple path nodes and path segments. Based on the global reference path information and status information of each AGV, the estimated time window for each AGV on the path node and path segment is determined. The estimated time window includes the path segment time window and the node time window. The path segment time window represents the time period taken by the AGV from entering the path segment to leaving the path segment. The node time window represents the time period during which the AGV is expected to arrive at and completely pass through the path node. Based on the estimated time window occupied by each AGV at each path node and path segment, determine whether there is spatiotemporal overlap; When spatiotemporal overlap is detected, the conflict type is determined based on the motion direction relationship of all AGVs corresponding to the spatiotemporal overlap; The conflict detection result is determined based on the conflict type, conflict spatial location, the identifiers of all AGVs involved in the conflict, and the estimated occupancy time window of the conflict spatial location corresponding to each AGV.

[0014] The beneficial effects of adopting the above-mentioned further scheme are: decomposing the global reference path information into path nodes and path segments, determining the estimated time window occupied by each AGV on it, accurately judging whether there is spatiotemporal overlap, and then determining the conflict type and obtaining the conflict detection result based on the movement direction relationship. It can proactively predict potential conflicts and classify them, providing a basis for subsequent intelligent invocation of multi-dimensional resolution strategy library for rapid decision-making, and ensuring the efficient, safe and smooth operation of AGV clusters in complex dynamic environments.

[0015] Furthermore, the central scheduling platform, when determining at least one target AGV based on the task request, the status information of each AGV, and a preset AGV capability evaluation model, specifically uses the following methods: The capacity value of each AGV is determined based on its battery power and average speed. The real-time load rate of each AGV is determined based on the current number of tasks for each AGV. The trust level of each AGV is determined based on its historical task completion status. Based on the task request, the capability values ​​of each AGV, the real-time load rate, the trust level, and the preset AGV capability evaluation model, at least one target AGV is determined.

[0016] The beneficial effects of adopting the above-mentioned further scheme are as follows: the capability value is determined based on the AGV's power consumption and average speed, the real-time load rate is determined based on the current number of tasks, the trust level is determined based on the historical task completion status, and the target AGV is determined by combining the task request and the AGV capability evaluation model. This enables intelligent task allocation for multi-AGV collaborative operation, introduces an intelligent screening mechanism based on indicators such as capability, trust level, and load, changes the task bidding from "broadcast" to "directed", and uses a multi-objective evaluation algorithm to reduce system communication volume and improve task allocation efficiency.

[0017] Furthermore, the central scheduling platform, when performing global path planning for each of the target AGVs based on the task request, the status information of each target AGV, and a preset path planning algorithm, specifically uses the following methods: Based on the task request, the status information of each target AGV, and the A* algorithm and / or Dijkstra's algorithm, global path planning is performed on each target AGV to obtain the global path of each target AGV; Key points are extracted from each of the global paths to obtain global reference path information for each target AGV, and the global reference path information includes a sequence of key points.

[0018] The beneficial effects of adopting the above-mentioned further scheme are: by using the A* algorithm and / or Dijkstra's algorithm in combination with the task request and the target AGV status information for global path planning, the optimal global path can be obtained; by extracting key points from the global path to form a key point sequence as global reference path information, the path information can be simplified, the computational load of subsequent local planning can be reduced, and the path planning efficiency can be improved.

[0019] Furthermore, the onboard intelligent unit, when performing real-time local optimization of the global reference path information based on the local environmental data to generate a local obstacle avoidance trajectory, specifically includes: Based on the dynamic window method and the local environment data, local real-time path planning is performed on the global reference path information. The evaluation function of the dynamic window method includes at least an orientation angle evaluation sub-function for measuring the difference between the direction of the trajectory end and the orientation of the current sub-target point, a distance evaluation sub-function for ensuring that the trajectory maintains a safe distance from obstacles, and a path deviation evaluation sub-function for guiding the AGV back to the global path. During the process of the target AGV traveling based on local real-time path planning, a key point is selected from the key point sequence as the current sub-target point based on the preset key point selection rules, and the distance between the target AGV and the current sub-target point is calculated. When the distance between the target AGV and the current sub-target point is less than the first set threshold, the current sub-target point is switched to the next key point selected in the key point sequence; When the target AGV detects a dynamic obstacle, it is guided to avoid the obstacle based on the distance evaluation sub-function, and after the target AGV avoids the obstacle, it is guided to smoothly return to the global path based on the path deviation evaluation sub-function. When the target AGV detects that the path is blocked or the deviation exceeds the second set threshold, it sends a replanning request to the central scheduling platform, triggering global path replanning.

[0020] The beneficial effects of adopting the above-mentioned further scheme are as follows: Local real-time path planning is performed based on the dynamic window method and local environmental data to obtain global reference path information. The orientation angle evaluation sub-function, distance evaluation sub-function, and path deviation evaluation sub-function in its evaluation function enable the target AGV to better navigate, avoid obstacles, and return to the global path during its journey. The current sub-target point is selected through preset key point selection rules and switched according to distance, allowing the target AGV to travel according to the planned key points. When dynamic obstacles are detected, obstacle avoidance is guided by the distance evaluation sub-function, and the path deviation evaluation sub-function guides the return to the global path, ensuring the safety of the target AGV and maintaining path planning. When path congestion or excessive deviation is detected, global path replanning is triggered, which can cope with complex environmental changes and improve the system's adaptability and reliability.

[0021] Secondly, this application provides a method for collaborative scheduling of unmanned vehicle clusters, employing the following technical solution: A method for collaborative scheduling of unmanned vehicle clusters includes: Obtain task requests for transporting stones and coal, as well as the status information of each AGV in the AGV cluster. The task requests include task location and task priority, and the status information includes location, battery level, average speed, and current task quantity. Based on the task request, the status information of each AGV, and the preset AGV capability evaluation model, at least one target AGV is determined; Based on the task request, the status information of each target AGV, and the preset path planning algorithm, global path planning is performed on each target AGV to obtain global reference path information for each target AGV. Each of the global reference path information is sent to the on-board intelligent unit corresponding to each target AGV, so that the target AGV can obtain the local environmental data of the target AGV in real time during the process of driving based on the global reference path information, perform local real-time optimization on the global reference path information based on the local environmental data, generate a local obstacle avoidance trajectory, and control the target AGV to drive according to the local obstacle avoidance trajectory.

[0022] Thirdly, this application provides an electronic device that adopts the following technical solution: An electronic device includes a memory and a processor, wherein the memory stores a computer program that can be loaded by the processor and execute the unmanned vehicle cluster cooperative scheduling method described in the second aspect.

[0023] Fourthly, this application provides a computer-readable storage medium, which adopts the following technical solution: A computer-readable storage medium storing a computer program that can be loaded by a processor and executed by the unmanned vehicle cluster cooperative scheduling method described in the second aspect.

[0024] Additional aspects and advantages of this application will be set forth in part in the description which follows, and will become apparent from the description or may be learned by practice of this application. Attached Figure Description

[0025] Figure 1 This is a schematic diagram of the structure of an unmanned vehicle cluster collaborative scheduling system provided in one embodiment of the present invention; Figure 2 This is a flowchart illustrating a method for collaborative scheduling of unmanned vehicle clusters according to an embodiment of the present invention. Figure 3 This is a schematic diagram of the structure of an electronic device provided in one embodiment of the present invention. Detailed Implementation

[0026] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0027] Furthermore, the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. Additionally, the character " / " in this article, unless otherwise specified, generally indicates that the preceding and following related objects have an "or" relationship.

[0028] like Figure 1 As shown, an unmanned vehicle cluster collaborative scheduling system 100 includes a central scheduling platform 101 and multiple on-board intelligent units 102 corresponding one-to-one with AGVs. The central scheduling platform 101 is used to obtain task requests for transporting stones and coal and the status information of each AGV in the AGV cluster. The task request includes the task location and task priority, and the status information includes the location, power, average speed and current number of tasks. The central scheduling platform 101 is used to determine at least one target AGV based on the task request, the status information of each AGV, and a preset AGV capability evaluation model. The central scheduling platform 101 is used to perform global path planning for each target AGV based on the task request, the status information of each target AGV and the preset path planning algorithm, to obtain global reference path information for each target AGV, and to send each global reference path information to the vehicle-mounted intelligent unit 102 corresponding to each target AGV. The on-board intelligent unit 102 corresponding to any of the target AGVs is used to acquire local environmental data of the target AGV in real time during the process of the target AGV driving based on the global reference path information, perform local real-time optimization of the global reference path information based on the local environmental data, generate a local obstacle avoidance trajectory, and control the target AGV to drive according to the local obstacle avoidance trajectory.

[0029] In this embodiment, the unmanned vehicle cluster collaborative scheduling system 100 also includes a unified interface module, which interfaces with power plant systems such as DCS, SIS, and MES. It obtains task requests for transporting coal and stone and the status information of each AGV by interacting with relevant power plant systems. It can also collect AGV status information in real time through sensors and other devices. For example, it can connect to the power plant's DCS system to obtain the coal mill's full coal bin signal, thereby receiving task requests; simultaneously, it obtains the AGV's position and power information through positioning and power sensors installed on the AGVs. Other data transmission methods and devices, such as wireless communication modules, can also be used to acquire information.

[0030] Subsequently, the central scheduling platform 101 determines at least one target AGV based on the task request, the status information of each AGV, and the preset AGV capability evaluation model. For example, it comprehensively evaluates the AGVs based on factors such as their capability value, real-time load rate, and trust level, and selects the most suitable AGV to execute the task.

[0031] Subsequently, the central dispatch platform 101 performs global path planning for each target AGV based on the task request, the status information of each target AGV, and a preset path planning algorithm. This yields global reference path information for each target AGV, which is then sent to the corresponding onboard intelligent unit 102. In this embodiment, path planning algorithms such as the A* algorithm or Dijkstra's algorithm can be used to generate the optimal global path. For example, the path planning module calculates the shortest path using the A* algorithm based on the task location and the AGV's current position, and extracts key points from the path to obtain global reference path information. Alternatively, the path can be dynamically adjusted by combining map information and real-time traffic conditions.

[0032] For any of the target AGVs corresponding to the onboard intelligent unit 102, local environmental data of the target AGV is acquired in real time during the AGV's movement based on global reference path information. In this embodiment, local environmental data, such as the position and distance of obstacles, can be acquired using sensors such as LiDAR and cameras. For example, a LiDAR can be installed on the AGV to scan the environment around the AGV in real time and acquire the distance and angle information of obstacles. Other types of sensors, such as ultrasonic sensors, can also be used.

[0033] Subsequently, the onboard intelligent unit 102 performs real-time local optimization of the global reference path information based on local environmental data, generates a local obstacle avoidance trajectory, and controls the target AGV to travel according to the local obstacle avoidance trajectory. For example, the onboard intelligent unit 102 adjusts the AGV's motor speed and steering angle according to the local obstacle avoidance trajectory, so that the AGV travels along the trajectory. The various modules of the onboard intelligent unit 102 cooperate with each other to realize local environmental perception and path optimization of the AGV during the driving process.

[0034] Optionally, the central scheduling platform 101, when determining at least one target AGV based on the task request, the status information of each AGV, and a preset AGV capability evaluation model, specifically uses the following methods: The capacity value of each AGV is determined based on its battery power and average speed. The real-time load rate of each AGV is determined based on the current number of tasks for each AGV. The trust level of each AGV is determined based on its historical task completion status. Based on the task request, the capability values ​​of each AGV, the real-time load rate, the trust level, and the preset AGV capability evaluation model, at least one target AGV is determined.

[0035] In this embodiment, the central scheduling platform 101 calculates the capability value of each AGV based on its current battery level and average speed, calculates the real-time load rate based on the current number of tasks, and calculates the trust level based on historical task completion data. Then, using a preset AGV capability evaluation model, it calculates the evaluation value for each AGV and selects the AGV with the highest evaluation value as the target AGV. Other evaluation models and algorithms, such as genetic algorithms, can also be used for task allocation.

[0036] Among them, Value_i= f(Load_i,Cap_i,Bel_i).

[0037] Where Load_i is the load factor, Cap_i is the capacity value, and Bel_i is the trust level.

[0038] Value_i= (W_l* (1 -Load_i))^α * (W_c*Cap_i)^β * (W_b*Bel_i)^γ.

[0039] The weighting coefficients (W_l, W_c, W_b) are used to adjust the relative importance of load factor, capability value, and trust level in the overall evaluation. For example, in scenarios where efficiency is maximized, the capability value weight W_c can be set higher; in scenarios with extremely high task reliability requirements, the trust level weight W_b should be set higher. All weights must satisfy the normalization condition (e.g., W_l + W_c + W_b = 1).

[0040] The exponents (α, β, γ) are used to perform a non-linear transformation on each indicator, thereby adjusting its "discrimination". For example, setting the trust index γ to be greater than 1 can amplify the advantages of high-trust AGVs, making the system more inclined to select vehicles with excellent historical performance.

[0041] Load_i, the real-time load rate, reflects the current busy level of the AGV. Load_i = (Current_Task_Number / Max_Task_Capacity).

[0042] Cap_i is an inherent capability value used to measure the AGV's ability to complete a certain type of task. It is calculated based on the AGV's operational measurement parameters (such as current battery level and recent average travel speed).

[0043] Bel_i trust score is used to quantify the historical reputation of an AGV in reliably completing a specific type of task. Bel_i trust score is dynamically updated based on historical task completion performance. Bel_i_new = (1-λ) * Bel_i_old + λ * (Success_Flag). Where λ is the learning rate, with Success_Flag = 1 for timely success and 0 for failure.

[0044] W_l, W_c, and W_b are weighting coefficients used to balance the relative importance of load, capability, and trust. The sum of the three is 1, determined according to the strategic objectives of the specific scenario.

[0045] α, β, and γ are exponential factors used to control the "sensitivity" or "discrimination" of each indicator. More refined nonlinear adjustments are made to further optimize the selection strategy.

[0046] Optionally, the central dispatch platform 101 is also used for: Based on the global reference path information and status information of each AGV, conflict detection is performed on each AGV to obtain the conflict detection result. Based on the conflict detection results and a pre-set traffic rule base, a conflict resolution strategy is generated for one or more AGVs among all the AGVs involved in the conflict detection results. The strategy includes at least one of speed adjustment, path replanning, or waiting at a safe point. The conflict resolution strategy is sent to the on-board intelligent unit 102 of the AGV corresponding to the conflict resolution strategy.

[0047] The vehicle-mounted intelligent unit 102 is also used to determine, based on the task request, whether the execution of the conflict resolution strategy by the AGV will affect the time for the AGV to execute the current task. If the execution of the conflict resolution strategy by the AGV will affect the time for the AGV to execute the current task, then an arbitration request is sent to the central scheduling platform 101. The central scheduling platform 101 is also used to make arbitration decisions based on the arbitration request and the task priority corresponding to the AGV.

[0048] In this embodiment, the central scheduling platform 101 generates conflict resolution strategies for one or more AGVs among all AGVs involved in the conflict detection results based on the conflict detection results and the preset traffic rule base. The strategies include at least one of speed adjustment, path replanning, or waiting at a safe point, and sends the conflict resolution strategies to the on-board intelligent unit 102 of the AGV corresponding to the conflict resolution strategy.

[0049] The central scheduling platform 101 selects an appropriate conflict resolution strategy based on the type of conflict and the specific circumstances. For example, when a reverse intersection conflict is detected, if the conflict can be avoided by slightly accelerating or decelerating to stagger the time window, the conflict resolution module will choose a speed adjustment strategy; if speed adjustment cannot resolve the conflict, the conflict resolution module will plan a slightly longer alternative path for the lower priority AGV, i.e., adopt a path replanning strategy; if there is a preset waiting area ahead of the conflict point, the conflict resolution module will have one AGV wait at the safe point until the other AGV has passed before proceeding.

[0050] The AGV evaluates the received conflict resolution strategies. If an updated strategy affects task completion time, it submits the case to the platform for arbitration, with the platform prioritizing task completion. If the tasks of the conflicting parties have a time conflict, the priority of the task of the coal mill with the heavier load is adjusted according to the coal mill's workload.

[0051] In this embodiment of the application, the central scheduling platform 101 is used to perform conflict detection on each AGV based on the global reference path information and status information of each AGV. When obtaining the conflict detection result, it is specifically used for: Based on the global reference path information of each AGV in the AGV cluster, each global reference path information is decomposed into multiple path nodes and path segments. Based on the global reference path information and status information of each AGV, the estimated time window for each AGV on the path node and path segment is determined. The estimated time window includes the path segment time window and the node time window. The path segment time window represents the time period taken by the AGV from entering the path segment to leaving the path segment. The node time window represents the time period during which the AGV is expected to arrive at and completely pass through the path node. Based on the estimated time window occupied by each AGV at each path node and path segment, determine whether there is spatiotemporal overlap; When spatiotemporal overlap is detected, the conflict type is determined based on the motion direction relationship of all AGVs corresponding to the spatiotemporal overlap; The conflict detection result is determined based on the conflict type, conflict spatial location, the identifiers of all AGVs involved in the conflict, and the estimated occupancy time window of the conflict spatial location corresponding to each AGV.

[0052] In the above implementation, the central scheduling platform 101 decomposes the global reference path information of the AGVs into multiple path nodes and path segments, calculates the estimated arrival and departure times of each AGV at these nodes and segments, forming time windows. Then, it compares the time windows of different AGVs to determine whether there is overlap. If there is overlap, a conflict is considered to have occurred. Other conflict detection methods, such as rule-based conflict detection methods, can also be used.

[0053] Optionally, the central scheduling platform 101, when performing global path planning for each of the target AGVs based on the task request, the status information of each target AGV, and a preset path planning algorithm, specifically uses the following: Based on the task request, the status information of each target AGV, and the A* algorithm and / or Dijkstra's algorithm, global path planning is performed on each target AGV to obtain the global path of each target AGV; Key points are extracted from each of the global paths to obtain global reference path information for each target AGV, and the global reference path information includes a sequence of key points.

[0054] By combining the A* algorithm and / or Dijkstra's algorithm with the task request and target AGV status information for global path planning, the optimal global path can be obtained. Extracting key points from the global path to form a key point sequence as global reference path information can simplify path information, reduce the computational load of subsequent local planning, and improve path planning efficiency.

[0055] Optionally, the onboard intelligent unit 102, when performing local real-time optimization of the global reference path information based on the local environmental data to generate a local obstacle avoidance trajectory, specifically performs the following: Based on the dynamic window method and the local environment data, local real-time path planning is performed on the global reference path information. The evaluation function of the dynamic window method includes at least an orientation angle evaluation sub-function for measuring the difference between the direction of the trajectory end and the orientation of the current sub-target point, a distance evaluation sub-function for ensuring that the trajectory maintains a safe distance from obstacles, and a path deviation evaluation sub-function for guiding the AGV back to the global path. During the process of the target AGV traveling based on local real-time path planning, a key point is selected from the key point sequence as the current sub-target point based on the preset key point selection rules, and the distance between the target AGV and the current sub-target point is calculated. When the distance between the target AGV and the current sub-target point is less than the first set threshold, the current sub-target point is switched to the next key point selected in the key point sequence; When the target AGV detects a dynamic obstacle, it is guided to avoid the obstacle based on the distance evaluation sub-function, and after the target AGV avoids the obstacle, it is guided to smoothly return to the global path based on the path deviation evaluation sub-function. When the target AGV detects that the path is blocked or the deviation exceeds the second set threshold, it sends a replanning request to the central scheduling platform 101, triggering global path replanning.

[0056] In this embodiment, key points are extracted after the global planning stage. For example, by determining whether three adjacent points in the path point set are collinear, redundant points are eliminated, forming an optimized path set composed of key turning points. These key points are the "sub-objectives" of the local path planning.

[0057] Local planners (such as DWA) use the next key point after the current key point as the current temporary target. When generating the trajectory using sampling velocity, the heading(v,w) sub-function in the DWA algorithm's evaluation function measures the difference between the direction of the trajectory's end and the orientation of the current sub-target point. While avoiding obstacles, the AGV is constantly being "pulled" towards the next key point on the global path.

[0058] When the distance between the AGV and the current sub-target is less than a set threshold, the sub-target will automatically switch to the next point in the key point sequence.

[0059] When a local adjustment "conflicts" with the global path (i.e., local obstacle avoidance behavior leads to a significant deviation), the system responds through the following mechanisms: Local adjustments take priority: When faced with suddenly appearing dynamic obstacles, the local planner has the highest priority to ensure safety. The dist(v,w) sub-function in the DWA's evaluation function ensures that the trajectory maintains a safe distance from the obstacle. At this time, the robot may temporarily deviate from the global path.

[0060] Traction to return to the global path: Once out of danger, the path(v,w) sub-function in the DWA evaluation function will guide the AGV to smoothly return to the global path, instead of driving directly to the sub-target point where it deviated, thus reducing unnecessary detours.

[0061] Global rerouting mechanism: This is the last resort for handling major conflicts. The system will trigger global path rerouting when the following conditions occur: Encountering a dead end: Local perception detects a fixed obstacle ahead that cannot be bypassed, blocking the global path.

[0062] Severe deviation: Due to continuous obstacle avoidance, the robot deviates too far from the global path, making it inefficient or even impossible to proceed according to the current sequence of key points.

[0063] After replanning, a new global path and key point sequence will be generated, and the local planner will continue navigation based on the new "roadbook".

[0064] The central scheduling platform 101 of this system determines the target AGV based on the task request and AGV status information, combined with the capability assessment model, and performs global path planning to provide global reference path information for the target AGV, thereby realizing intelligent task allocation and path planning for multi-AGV collaborative operations. The on-board intelligent unit 102 can perform local real-time optimization of the global reference path information based on local environmental data, generate local obstacle avoidance trajectories, improve the system's adaptability and safety in complex environments, and enhance the overall system efficiency and resource utilization. Figure 2 This is a flowchart illustrating a collaborative scheduling method for unmanned vehicle clusters provided in an embodiment of this application.

[0065] like Figure 2 As shown, a method for collaborative scheduling of unmanned vehicle clusters mainly includes: S201, Obtain the task request for transporting stones and coal and the status information of each AGV in the AGV cluster. The task request includes the task location and task priority, and the status information includes the location, battery level, average speed, and current task quantity. S202, based on the task request, the status information of each AGV, and the preset AGV capability evaluation model, determine at least one target AGV; S203, based on the task request, the status information of each target AGV and the preset path planning algorithm, perform global path planning for each target AGV to obtain global reference path information for each target AGV; S204, each of the global reference path information is sent to the on-board intelligent unit corresponding to each of the target AGVs, so that the target AGV can obtain the local environmental data of the target AGV in real time during the process of driving based on the global reference path information, perform local real-time optimization on the global reference path information based on the local environmental data, generate a local obstacle avoidance trajectory, and control the target AGV to drive according to the local obstacle avoidance trajectory.

[0066] Figure 3 This is a structural block diagram of an electronic device 300 according to an embodiment of this application.

[0067] like Figure 3 As shown, the electronic device 300 includes a processor 301 and a memory 302, and may further include one or more of an information input / output (I / O) interface 303, a communication component 304, and a communication bus 305.

[0068] The processor 301 controls the overall operation of the electronic device 300 to complete all or part of the steps in the aforementioned unmanned vehicle cluster collaborative scheduling method. The memory 302 stores various types of data to support the operation of the electronic device 300. This data may include, for example, instructions for any application or method operating on the electronic device 300, as well as application-related data. The memory 302 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as one or more of Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read-Only Memory (EPROM), Programmable Read-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0069] I / O interface 303 provides an interface between processor 301 and other interface modules, such as keyboards, mice, and buttons. These buttons can be virtual or physical. Communication component 304 is used to test wired or wireless communication between electronic device 300 and other devices. Wireless communication includes Wi-Fi, Bluetooth, Near Field Communication (NFC), 2G, 3G, or 4G, or a combination thereof. Therefore, the corresponding communication component 304 may include a Wi-Fi component, a Bluetooth component, and an NFC component.

[0070] The communication bus 305 may include a path for transmitting information between the aforementioned components. The communication bus 305 may be a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus, etc. The communication bus 305 may be divided into an address bus, a data bus, a control bus, etc.

[0071] The electronic device 300 may be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field-programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to execute the unmanned vehicle cluster cooperative scheduling method given in the above embodiments.

[0072] The following describes the computer-readable storage medium provided in the embodiments of this application. The computer-readable storage medium described below can be referred to in correspondence with the unmanned vehicle cluster collaborative scheduling method described above.

[0073] This application also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the above-described unmanned vehicle cluster collaborative scheduling method.

[0074] The computer-readable storage medium may include various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0075] The terms “comprising,” “including,” or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0076] The above description is merely a preferred embodiment of this application and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of this application is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the foregoing application concept. For example, technical solutions formed by substituting the above features with (but not limited to) technical features with similar functions claimed in this application.

Claims

1. A collaborative scheduling system for unmanned vehicle clusters, characterized in that, This includes a central dispatch platform and multiple onboard intelligent units that correspond one-to-one with each AGV; The central scheduling platform is used to obtain task requests for transporting stones and coal, as well as the status information of each AGV in the AGV cluster. The task request includes the task location and task priority, and the status information includes the location, battery level, average speed, and current number of tasks. The central scheduling platform is used to determine at least one target AGV based on the task request, the status information of each AGV, and a preset AGV capability evaluation model. The central scheduling platform is used to perform global path planning for each target AGV based on the task request, the status information of each target AGV and the preset path planning algorithm, to obtain global reference path information for each target AGV, and to send each global reference path information to the on-board intelligent unit corresponding to each target AGV. The on-board intelligent unit corresponding to any of the target AGVs is used to acquire local environmental data of the target AGV in real time during the process of the target AGV driving based on the global reference path information, perform local real-time optimization of the global reference path information based on the local environmental data, generate a local obstacle avoidance trajectory, and control the target AGV to drive according to the local obstacle avoidance trajectory.

2. The unmanned vehicle cluster collaborative scheduling system according to claim 1, characterized in that, The central dispatch platform is also used for: Based on the global reference path information and status information of each AGV, conflict detection is performed on each AGV to obtain the conflict detection result. Based on the conflict detection results and a pre-set traffic rule base, a conflict resolution strategy is generated for one or more AGVs among all the AGVs involved in the conflict detection results. The strategy includes at least one of speed adjustment, path replanning, or waiting at a safe point. The conflict resolution strategy is sent to the onboard intelligent unit of the AGV corresponding to the conflict resolution strategy.

3. The unmanned vehicle cluster collaborative scheduling system according to claim 2, characterized in that, The onboard intelligent unit is also used to determine, based on the task request, whether the execution of the conflict resolution strategy by the AGV will affect the time for the AGV to execute the current task. If the execution of the conflict resolution strategy by the AGV will affect the time for the AGV to execute the current task, then an arbitration request is sent to the central scheduling platform. The central scheduling platform is also used to make arbitration decisions based on the arbitration request and the task priority corresponding to the AGV.

4. The unmanned vehicle cluster collaborative scheduling system according to claim 2, characterized in that, The central scheduling platform is used to perform conflict detection on each AGV based on the global reference path information and status information of each AGV. When obtaining the conflict detection result, it is specifically used for: Based on the global reference path information of each AGV in the AGV cluster, each global reference path information is decomposed into multiple path nodes and path segments. Based on the global reference path information and status information of each AGV, the estimated time window for each AGV on the path node and path segment is determined. The estimated time window includes the path segment time window and the node time window. The path segment time window represents the time period taken by the AGV from entering the path segment to leaving the path segment. The node time window represents the time period during which the AGV is expected to arrive at and completely pass through the path node. Based on the estimated time window occupied by each AGV at each path node and path segment, determine whether there is spatiotemporal overlap; When spatiotemporal overlap is detected, the conflict type is determined based on the motion direction relationship of all AGVs corresponding to the spatiotemporal overlap; The conflict detection result is determined based on the conflict type, conflict spatial location, the identifiers of all AGVs involved in the conflict, and the estimated occupancy time window of the conflict spatial location corresponding to each AGV.

5. The unmanned vehicle cluster collaborative scheduling system according to claim 1, characterized in that, The central scheduling platform, when determining at least one target AGV based on the task request, the status information of each AGV, and a preset AGV capability evaluation model, specifically uses the following: The capacity value of each AGV is determined based on its battery power and average speed. The real-time load rate of each AGV is determined based on the current number of tasks for each AGV. The trust level of each AGV is determined based on its historical task completion status. Based on the task request, the capability values ​​of each AGV, the real-time load rate, the trust level, and the preset AGV capability evaluation model, at least one target AGV is determined.

6. The unmanned vehicle cluster collaborative scheduling system according to claim 1, characterized in that, The central scheduling platform, when performing global path planning for each target AGV based on the task request, the status information of each target AGV, and a preset path planning algorithm, is specifically used for: Based on the task request, the status information of each target AGV, and the A* algorithm and / or Dijkstra's algorithm, global path planning is performed on each target AGV to obtain the global path of each target AGV; Key points are extracted from each of the global paths to obtain global reference path information for each target AGV, and the global reference path information includes a sequence of key points.

7. The unmanned vehicle cluster collaborative scheduling system according to claim 6, characterized in that... The vehicle-mounted intelligent unit, when performing real-time local optimization of the global reference path information based on the local environmental data to generate a local obstacle avoidance trajectory, specifically includes: Based on the dynamic window method and the local environment data, local real-time path planning is performed on the global reference path information. The evaluation function of the dynamic window method includes at least an orientation angle evaluation sub-function for measuring the difference between the direction of the trajectory end and the orientation of the current sub-target point, a distance evaluation sub-function for ensuring that the trajectory maintains a safe distance from obstacles, and a path deviation evaluation sub-function for guiding the AGV back to the global path. During the process of the target AGV driving based on local real-time path planning, a key point is selected from the key point sequence as the current sub-target point based on the preset key point selection rules, and the distance between the target AGV and the current sub-target point is calculated. When the distance between the target AGV and the current sub-target point is less than the first set threshold, the current sub-target point is switched to the next key point selected in the key point sequence; When the target AGV detects a dynamic obstacle, it is guided to avoid the obstacle based on the distance evaluation sub-function, and after the target AGV avoids the obstacle, it is guided to smoothly return to the global path based on the path deviation evaluation sub-function. When the target AGV detects that the path is blocked or the deviation exceeds the second set threshold, it sends a replanning request to the central scheduling platform, triggering global path replanning.

8. A method for collaborative scheduling of unmanned vehicle clusters, characterized in that, include: Obtain task requests for transporting stones and coal, as well as the status information of each AGV in the AGV cluster. The task requests include task location and task priority, and the status information includes location, battery level, average speed, and current task quantity. Based on the task request, the status information of each AGV, and the preset AGV capability evaluation model, at least one target AGV is determined; Based on the task request, the status information of each target AGV, and the preset path planning algorithm, global path planning is performed on each target AGV to obtain global reference path information for each target AGV. Each of the global reference path information is sent to the on-board intelligent unit corresponding to each target AGV, so that the target AGV can obtain the local environmental data of the target AGV in real time during the process of driving based on the global reference path information, perform local real-time optimization on the global reference path information based on the local environmental data, generate a local obstacle avoidance trajectory, and control the target AGV to drive according to the local obstacle avoidance trajectory.

9. An electronic device, characterized in that, Includes a processor, which is coupled to a memory; The processor is configured to execute a computer program stored in the memory, so that the electronic device performs the method as described in claim 8.

10. A computer-readable storage medium, characterized in that, It includes a computer program or instructions that, when run on a computer, cause the computer to perform the method as described in claim 8.

Images