Scheduling method and system for intelligent mobile robot cluster

Abstract

The invention relates to the technical field of intelligent mobile robots, and particularly discloses a scheduling method for an intelligent mobile robot cluster, and the method comprises the steps: constructing a task priority evaluation model, and calculating the priority of each task; establishing a robot state evaluation index system, and calculating the state score of the robot according to the actual measurement information of the robot; calculating the crowdedness degree of each path from the robot to the task place; performing task allocation by adopting an improved Hungary algorithm, establishing a task allocation cost matrix, and introducing a penalty function; and performing path planning in combination with the environment information, calculating the comprehensive cost of each path, and selecting the path with the minimum comprehensive cost. Through reasonable task classification, priority setting and an optimized task allocation algorithm, high-priority tasks can be ensured to be processed in time, meanwhile, the tasks are allocated among the robots in a balanced mode, the situation that loads of the robots are unbalanced is avoided, and therefore the task execution efficiency of the whole robot cluster is effectively improved.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent mobile robots, and specifically relates to a scheduling method and system for an intelligent mobile robot cluster. Background Art

[0002] With the rapid development of fields such as intelligent manufacturing and logistics automation, intelligent mobile robot clusters are increasingly widely used in various scenarios. Intelligent mobile robot clusters can cooperate to complete complex tasks, such as cargo handling, warehouse management, environmental monitoring, etc. Compared with single robots, they have higher efficiency and stronger adaptability.

[0003] In the wide range of application scenarios of intelligent mobile robot clusters, path planning, as a core technical link, plays a crucial role in the efficient operation of robots. From material handling in industrial production to cargo distribution in logistics warehousing, and then to inspection operations in complex environments, robots need accurate and efficient path planning to achieve task goals. However, current path planning technologies face many challenges, which seriously restrict the performance of intelligent mobile robot clusters.

[0004] In terms of adaptability to dynamic environments, the operating environments of robots are often complex and changeable. Taking the warehouse logistics scenario as an example, the frequent handling of goods will cause the shelf layout to change at any time. Newly added cargo stacking areas or temporary obstacles will render the originally planned paths invalid. Traditional path planning algorithms usually calculate paths based on static maps and are difficult to perceive and respond to these environmental changes in real time. When encountering sudden obstacles, robots may get stuck and wait for re-planning of paths, which not only wastes time but may also cause the entire operation process to be interrupted, reducing the operating efficiency and stability of the system.

[0005] It is also a major problem that it is difficult to guarantee the global optimality of path planning. When some algorithms plan paths, they overly focus on local optimal solutions, only considering the environmental information near the current position to select paths and ignoring global environmental factors and the overall requirements of tasks. For example, in a multi-robot cooperative operation scenario, when each robot independently plans a path, it may choose a route that seems locally optimal but conflicts with the paths of other robots, resulting in robots waiting for and avoiding each other, reducing the overall operation efficiency. Moreover, these algorithms do not fully consider the priorities of tasks and the states of robots, making high-priority tasks unable to be processed in a timely manner, affecting the task execution effect of the entire system. Summary of the Invention

[0006] The purpose of the present invention is to provide a scheduling method and system for an intelligent mobile robot cluster to solve the above technical problems.

[0007] The purpose of the present invention can be achieved by the following technical solutions:

[0008] A scheduling method for a cluster of intelligent mobile robots, comprising the following steps:

[0009] Obtain the corresponding timeliness score and impact degree score according to the task content, construct a task priority evaluation model, and calculate the priority of each task;

[0010] Establish a robot state evaluation index system, and calculate the state score of the robot according to the measured information of the robot;

[0011] Obtain the environmental information of the working location where the robot is located, perform real-time analysis on the environmental information, and calculate the congestion degree of each path for the robot to travel to the task location;

[0012] Based on the task priority, robot state, and environmental information, use an improved Hungarian algorithm for task allocation, establish a task allocation cost matrix, and introduce a penalty function;

[0013] Plan the optimal path for each robot assigned with a task. Combining the environmental information, use the Dijkstra algorithm or A* algorithm for path planning, calculate the comprehensive cost of each path, and select the path with the minimum comprehensive cost.

[0014] As a further solution of the present invention: the steps of constructing the task priority evaluation model and calculating the priority of each task include:

[0015] The formula of the task priority evaluation model is as follows:

[0016] P i = αT i + βB i ;

[0017] Among them, P i represents the priority of task i, T i represents the timeliness score of task i, B i represents the impact degree score of task i on the overall business process, α and β are weight coefficients, and α + β = 1, α, β ∈ [0, 1].

[0018] As a further solution of the present invention: the steps of establishing the robot state evaluation index system and calculating the state score of the robot according to the measured information of the robot include:

[0019] Calculate the state score S j of robot j, and the calculation formula is as follows:

[0020]

[0021] E j is the remaining battery ratio of robot j, Fj is the number of failures of robot j, L j is the utilization rate of the load capacity of robot j, and γ, δ, and θ are weight coefficients, and γ + δ + θ = 1, γ, δ, θ ∈ [0, 1].

[0022] As a further solution of the present invention: Calculate the congestion degree C k of path k, and the calculation formula is as follows:

[0023] C k = n k / N k ;

[0024] where n k is the number of robots on path k per unit time, and N k is the maximum number of robots that path k can carry.

[0025] As a further solution of the present invention: In the task assignment cost matrix M, the element M ij is calculated by the following formula:

[0026]

[0027] where D ij represents the distance between task i and robot j;

[0028] Introduce a penalty function Penalty ij = μΔL ij + vΔE ij , and the adjusted element of the task assignment cost matrix is:

[0029] M' ij = M ij + Penalty ij ;

[0030] where ΔL ij is the expected load increase of robot j after accepting task i, and ΔE ij is the expected power consumption of robot j after accepting task i, and μ and ν are penalty coefficients.

[0031] As a further solution of the present invention: Use the Dijkstra algorithm or A* algorithm, etc. for path planning, and calculate the comprehensive cost Cost k of path k through the formula. The calculation formula of the comprehensive cost Cost k is:

[0032] Cost k = L k *(1 + ωC k );

[0033] Among them, L k is the length of path k, and ω is the congestion impact coefficient;

[0034] And select the path with the minimum comprehensive cost.

[0035] As a further solution of the present invention: when the robot encounters a failure or other emergencies, it sends a distress signal; according to the status scores and task priorities of other robots, it reallocates the robots to execute the current tasks.

[0036] A scheduling system for an intelligent mobile robot cluster, comprising:

[0037] Task grabbing and preprocessing module: obtain the corresponding timeliness score and impact degree score according to the task content, construct a task priority evaluation model, and calculate the priorities of each task;

[0038] Robot comprehensive evaluation module: establish a robot status evaluation index system, and calculate the status score of the robot according to the measured information of the robot;

[0039] Environment monitoring module: obtain the environmental information of the working location where the robot is located, perform real-time analysis on the environmental information, and calculate the congestion degree of each path for the robot to travel to the task location;

[0040] Path planning module: based on task priorities, robot status, and environmental information, use an improved Hungarian algorithm for task allocation, establish a task allocation cost matrix, and introduce a penalty function;

[0041] Plan the optimal path for each robot assigned a task, combine the environmental information, use the Dijkstra algorithm or A* algorithm for path planning, calculate the comprehensive cost of each path, and select the path with the minimum comprehensive cost.

[0042] The beneficial effects of the present invention: through reasonable task classification, priority setting, and optimized task allocation algorithms, it can ensure that high-priority tasks are processed in a timely manner, and at the same time evenly distribute tasks among robots, avoiding the situation of uneven robot loads, thereby effectively improving the task execution efficiency of the entire robot cluster.

[0043] Real-time environmental information collection and analysis, as well as dynamic path planning and adjustment mechanisms, enable the robot cluster to quickly respond to various changes in the working environment, such as the appearance of obstacles and path congestion, ensuring that the robots can successfully complete tasks and improving the operation stability and reliability of the robot cluster in complex environments. Brief Description of the Drawings

[0044] The present invention will be further described below with reference to the accompanying drawings.

[0045] Figure 1 It is a schematic flow chart of a scheduling method for an intelligent mobile robot cluster according to the present invention. Specific embodiments

[0046] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0047] Please refer to Figure 1 As shown, the present invention is a scheduling method for an intelligent mobile robot cluster, specifically including:

[0048] Task classification and priority setting:

[0049] According to factors such as the type, urgency, and required resources of the task, the tasks are divided into different categories, and priorities are set for each task. A task priority evaluation model is established, comprehensively considering factors such as the timeliness of the task and the degree of impact on the overall business process. Let the timeliness score of task i be T i , and the degree of impact on the overall business process be scored as B i , then the priority P i of task i can be calculated by the following formula:

[0050] P i = αT i + βB i ;

[0051] Among them, α and β are weight coefficients, and α + β = 1, α, β ∈ [0, 1], which can be adjusted according to the actual application scenario.

[0052] Suppose in this logistics warehouse scenario, for urgent order delivery tasks, due to their extremely high requirements for timeliness, if they cannot be completed on time, it will lead to customer complaints or even economic compensation. Therefore, the timeliness score T i of such tasks will be set relatively high, for example, 0.9. At the same time, the degree of impact of such tasks on the overall business process is also relatively large, because delayed delivery may affect subsequent production or sales links. Therefore, the impact degree score B i can be set to 0.8.

[0053] For ordinary order delivery tasks, the requirements for timeliness are relatively low, and the timeliness score T iIt can be set to 0.6, and the impact score on the overall business process is 0.5. For the warehouse inspection task, which is mainly to ensure the safety of the warehouse environment and the normal storage of goods, the timeliness requirement is not high. The timeliness score Ti is 0.2, and the impact score Bi on the overall business process is 0.3;

[0054] Calculate the task priority according to the formula Pi = αTi + βBi. Assume that in this logistics warehouse scenario, α = 0.6 and β = 0.4. For the urgent order delivery task, its priority P i = 0.6×0.9 + 0.4×0.8 = 0.86; for the general order delivery task, the priority P i = 0.6×0.6 + 0.4×0.5 = 0.56; for the warehouse inspection task, the priority P i = 0.6×0.2 + 0.4×0.3 = 0.24.

[0055] The timeliness score mainly measures the urgency of the task in terms of time and can be calculated in the following ways in the present invention:

[0056] 1. Linear scoring based on the deadline:

[0057] If the task has a clear deadline t deadline , and the current time is t now , a maximum allowable time difference T max (for example, the maximum duration from the task release to the deadline) can be set. Then the timeliness score T i can be calculated by the following formula:

[0058] T i = 1 - (t deadline - t now ) / T max ;

[0059] When t now is close to t deadline , T i approaches 1, indicating high task timeliness;

[0060] When t now is much less than t deadline , T i approaches 0, indicating low task timeliness;

[0061] 2. Phase scoring:

[0062] Divide the time range of the task into different phases, and assign different score values to each phase. For example, for a task, if the time to the deadline is within 1 hour, T i = 1; within 1 - 3 hours, T i= 0.7; within 3 - 6 hours, T i = 0.4; after more than 6 hours, T i = 0.1.

[0063] 3. Scoring based on delay cost:

[0064] Consider the cost C delay brought about by task delay, and the cost C normal of normally completing the task.

[0065] The timeliness score T i can be expressed as:

[0066] T i = C delay / (C delay + C normal );

[0067] The higher the delay cost, the higher the timeliness score.

[0068] The impact degree score measures the importance and influence of a task on the overall business process and can be calculated in the following ways in the present invention:

[0069] 1. Scoring based on business process nodes:

[0070] Divide the business process into multiple key nodes, and each task corresponds to one or more nodes. If a task affects a core business node, assign a higher score; if it affects a non-core node, assign a lower score. For example, in a logistics business process, the loading, unloading, and transportation of goods are core nodes, while the cleaning of the warehouse is a non-core node. It can be set that for core node tasks, B i = 0.8, and for non-core node tasks, B i = 0.2.

[0071] 2. Scoring based on task correlation:

[0072] Consider the degree of correlation between a task and other tasks. If the completion or non-completion of a task directly affects the progress of multiple other tasks, then its impact degree score is higher. The score can be calculated by calculating the number of associated tasks n related of the task and the total number N of all tasks:

[0073] B i = n related / N;

[0074] 3. Scoring based on business loss:

[0075] Evaluate the business loss L loss caused by the non-completion or delay of the task, and the total value L total. Impact degree score: B i It can be expressed as:

[0076] B i = L loss / L total ;

[0077] The greater the business loss, the higher the impact degree score.

[0078] Robot status monitoring and evaluation:

[0079] Real-time monitor information such as the battery level, position, running speed, load capacity, and fault status of the robot. Establish an index system for robot status evaluation to quantitatively evaluate the health status of the robot. Let the remaining battery percentage of robot j be E j , the number of faults be F j , and the load capacity utilization rate be L j . Then the status score S j of robot j can be calculated by the following formula:

[0080]

[0081] where γ, δ, and θ are weight coefficients, and γ + δ + θ = 1, γ, δ, θ ∈ [0, 1], which can be adjusted according to the actual situation.

[0082] Environmental information collection and analysis:

[0083] Use sensors such as lidar and cameras to collect information on the robot's working environment, including map information, obstacle distribution, path conditions, etc. Analyze the collected environmental information in real time to identify potential dangerous areas and path congestion points. Let the congestion degree of path k be C k , which can be calculated by the number of robots n k on this path per unit time and the maximum number of robots N k that the path can carry:

[0084] C k = n k / N k ;

[0085] Task allocation algorithm:

[0086] Based on task priorities, robot status, and environmental information, use an improved Hungarian algorithm for task allocation. When allocating tasks, give priority to allocating high-priority tasks to robots with high status scores and close distances to the task locations. Let the distance between task i and robot j be D ij . Establish a task allocation cost matrix M, where the element M ij can be expressed as:

[0087]

[0088] Based on the Hungarian algorithm, a penalty function Penalty is introduced to penalize the situations where the robot may have an excessive load or rapid power consumption after task allocation. Let the expected increase in load of robot j after accepting task i be ΔL ij , and the expected power consumption be ΔE ij . Then the penalty function is:

[0089] Penalty ij = μΔL ij + vΔE ij ;

[0090] Among them, μ and ν are penalty coefficients, which can be adjusted according to the actual situation. The elements of the adjusted task allocation cost matrix are:

[0091] M′ ij = M ij + Penalty ij .

[0092] Path planning and dynamic adjustment:

[0093] Plan the optimal path for each robot assigned a task. Combining environmental information, use algorithms such as Dijkstra's algorithm or A* algorithm for path planning. Let the length of path k be L k , and the congestion degree be C k . Then the comprehensive cost Cost k of path k can be expressed as:

[0094] Cost k = L k *(1 + ωC k );

[0095] Among them, ω is the congestion influence coefficient, which can be adjusted according to the actual situation.

[0096] Energy management strategy:

[0097] According to the power status and task requirements of the robot, formulate a reasonable energy management strategy. Establish a power prediction model, and based on the current power E j0 of the robot, running speed v j , load condition L j and remaining task distance d j and other factors, predict the remaining power situation E j1 of the robot after completing the task. Let the power consumption coefficient per unit distance be k E . Then the power prediction formula is:

[0098] Ej1 = E j0 - k E × d j × (1 + λL j );

[0099] where λ is the influence coefficient of the load on the power consumption and can be adjusted according to the actual situation.

[0100] When the power E of the robot j1 is lower than the set threshold E th , it is preferentially arranged to return to the charging point for charging, and at the same time, the task it is currently undertaking is reassigned to other robots with sufficient power.

[0101] Communication and cooperation mechanism:

[0102] Build an efficient communication network to ensure real-time and stable data transmission between robots and between robots and the scheduling system. Formulate cooperation rules between robots. When multiple robots complete a task together, clarify their respective responsibilities and divisions of labor.

[0103] During the process of the robot executing the task, monitor the environmental changes in real time. If it is found that the original path is congested or there are new obstacles, promptly adopt a dynamic path adjustment algorithm to re-plan the path for the robot.

[0104] The above has described a specific embodiment of the present invention in detail, but the content described is only the preferred embodiment of the present invention and cannot be considered as limiting the scope of implementation of the present invention. All equivalent changes and improvements made according to the scope of the application of the present invention should still fall within the scope covered by the patent of the present invention.

Claims

1. A scheduling method for a cluster of intelligent mobile robots, characterized in that, It includes the following steps: Obtain the corresponding timeliness score and impact degree score according to the task content, construct a task priority evaluation model, and calculate the priority of each task; Establish a robot state evaluation index system, and calculate the state score of the robot according to the measured information of the robot; Obtain the environmental information of the working location where the robot is located, perform real-time analysis on the environmental information, and calculate the congestion degree of each path for the robot to travel to the task location; Based on the task priority, robot state, and environmental information, use an improved Hungarian algorithm for task allocation, establish a task allocation cost matrix, and introduce a penalty function; Plan the optimal path for each robot assigned with a task. Combining with the environmental information, use the Dijkstra algorithm or A* algorithm for path planning, calculate the comprehensive cost of each path, and select the path with the minimum comprehensive cost.

2. The scheduling method for an intelligent mobile robot cluster according to claim 1, characterized in that, The steps of constructing the task priority evaluation model and calculating the priority of each task include: The formula of the task priority evaluation model is as follows: P i = αT i + βB i ; Among them, P i represents the priority of task i, T i represents the timeliness score of task i, B i represents the impact degree score of task i on the overall business process. α and β are weight coefficients, and α + β = 1, α, β ∈ [0, 1].

3. A scheduling method for an intelligent mobile robot cluster according to claim 2, characterized in that, The steps of establishing the robot state evaluation index system and calculating the state score of the robot according to the measured information of the robot include: The status score S of the computer robot j j , and the calculation formula is as follows: E j is the remaining battery percentage of robot j, F j is the number of failures of robot j, L j is the utilization rate of the load capacity of robot j, and γ, δ, and θ are weight coefficients, and γ + δ + θ = 1, γ, δ, θ ∈ [0, 1].

4. A scheduling method for an intelligent mobile robot cluster according to claim 3, characterized in that Calculate the congestion level C of path k k The calculation formula is as follows: C k = n k / N k ; where n k is the number of robots on path k per unit time, and N k is the maximum number of robots that path k can carry.

5. A scheduling method for an intelligent mobile robot cluster according to claim 4, characterized in that In the task assignment cost matrix M, the element M ij is calculated by the following formula: Among them, D ij represents the distance between task i and robot j; Introduce the penalty function Penalty ij = μΔL ij + vΔE ij , and the elements of the adjusted task assignment cost matrix are as follows: M′ ij = M ij + Penalty ij ; where, ΔL ij is the expected load increase of robot j after receiving task i, and ΔE ij is the expected power consumption of robot j after receiving task i, and μ and ν are penalty coefficients.

6. The scheduling method for an intelligent mobile robot cluster according to claim 5, characterized in that Path planning is carried out using Dijkstra's algorithm or A* algorithm, etc., and the comprehensive cost Cost of path k is calculated through formulas k , the comprehensive cost Cost k The calculation formula is as follows: Cost k = L k *(1 + ωC k ) Among them, L k is the length of path k, and ω is the congestion impact coefficient; And select the path with the minimum comprehensive cost.

7. A scheduling method for an intelligent mobile robot cluster according to claim 1, characterized in that, When the robot encounters a failure or other emergency, send a distress signal; re-allocate the robot to execute the current task according to the state score and task priority of other robots.

8. A scheduling system for a cluster of intelligent mobile robots, characterized in that, It includes: Task grabbing and preprocessing module: Obtain the corresponding timeliness score and impact degree score according to the task content, construct a task priority evaluation model, and calculate the priority of each task; Robot comprehensive evaluation module: Establish a robot state evaluation index system, and calculate the state score of the robot according to the measured information of the robot; Environmental monitoring module: Obtain the environmental information of the working location where the robot is located, perform real-time analysis on the environmental information, and calculate the congestion degree of each path for the robot to travel to the task location; Path planning module: Based on the task priority, robot state, and environmental information, use an improved Hungarian algorithm for task allocation, establish a task allocation cost matrix, and introduce a penalty function; Plan the optimal path for each robot assigned with a task. Combining with the environmental information, use the Dijkstra algorithm or A* algorithm for path planning, calculate the comprehensive cost of each path, and select the path with the minimum comprehensive cost.

Images