A method for line determination of a robot swarm

Abstract

The application discloses a kind of line determination methods of robot cluster, the method includes: step S1: receiving global job task, the global job task at least includes target area;Step S2: based on global map and the position information and state information of each robot in robot cluster, determine multiple task robots for executing the global job task;Step S3: based on the feasible path in the target area, the target area is divided into multiple sub-regions equal to the number of the task robot;Step S4: based on distance matching rule, the multiple sub-regions and the multiple task robots are one-to-one corresponding;Step S5: for each the task robot plans its corresponding sub-region in the job route.The method can realize job area dynamic division, robot intelligent screening and accurate determination of quantity, job path autonomous planning.

Description

Technical Field

[0001] This invention relates to the field of robotics, and in particular to a method for determining the route of a robot swarm. Background Technology

[0002] In the field of robot swarm operations, such as sanitation cleaning, park inspection, and warehouse sorting, a zoned operation scheme is often adopted to improve operational efficiency. In existing technologies, robot swarms need to pre-define a fixed overall operation area, and then divide it into multiple sub-areas with fixed boundaries. Each sub-area is equipped with a dedicated robot and a fixed operation path is preset. The robot can only complete repetitive tasks within a fixed range according to preset instructions.

[0003] However, this solution has obvious drawbacks when facing dynamic operation requirements: when the overall operation area changes, the size of the sub-area and the robot operation path need to be replanned manually, which is cumbersome; at the same time, it is not possible to quickly and reasonably select suitable robots for the changed tasks, and it is also difficult to determine the required number of robots, resulting in low operation efficiency and poor adaptability.

[0004] Therefore, there is an urgent need for a method to determine the robot swarm route that can achieve dynamic division of the work area, intelligent selection and precise determination of the number of robots, and autonomous planning of the work path, in order to overcome the shortcomings of the existing technology. Summary of the Invention

[0005] The purpose of this invention is to provide a method for determining the route of a robot swarm, which addresses the problems mentioned above. This method enables dynamic division of the work area, intelligent selection and precise determination of the number of robots, and autonomous planning of the work path.

[0006] The technical solution adopted in this invention is as follows: A method for determining the route of a robot swarm, the method comprising: Step S1: Receive a global job task, wherein the global job task contains at least a target area; Step S2: Based on the global map and the position and status information of each robot in the robot cluster, determine multiple task robots to perform the global task; Step S3: Based on the feasible paths within the target area, divide the target area into multiple sub-regions equal to the number of task robots; Step S4: Based on the distance matching rules, map each of the multiple sub-regions to one of the multiple task robots; Step S5: Plan the operation route in the corresponding sub-region for each task robot.

[0007] Furthermore, step S1 specifically includes: Step S1.1: Receive task configuration information from the user client; Step S1.2: Parse the task configuration information to generate structured global task data, wherein the global task data includes at least the range information of the target area; Step S1.3: Verify the integrity and validity of the global task data. If the verification passes, store the global task data and generate a global task based on the verified global task data for subsequent steps.

[0008] Furthermore, step S2 includes: Step S2.1: Obtain the global map and extract the target area, the key points contained in the target area, and the position and status information of each robot in the robot cluster from it; Step S2.2: For each robot in the robot cluster, based on its location information, the key points, and the obstacle areas marked in the global map, plan a feasible path and calculate the average reachable distance from the robot to each key point in the target area; Step S2.3: Prioritize each robot in the robot cluster based on multi-dimensional evaluation indicators to obtain priority ranking results. The multi-dimensional evaluation indicators include at least average reachability distance and status information. Combined with the workload of the target area, the robot's work efficiency and the work time requirements of the global task, determine the final number of robots participating in the operation. Step S2.4: Based on the determined number of robots and priority ranking results, select a corresponding number of robots from the robot cluster as the task robots.

[0009] Furthermore, in step S2.1, the global map is marked with obstacle areas and passable areas; the key points are the necessary passage points for the robot to enter the target area or the core operation nodes within the target area; the location information is the real-time coordinates of each robot, and the status information includes at least one of the following: the remaining energy of each robot, the current task status, and the path terrain adaptation information. In step S2.2, the feasible path is the obstacle avoidance and detour path planned by the robot from its current position to each key point in the target area, with the obstacle area as the constraint condition; the average reachable distance is the arithmetic mean of the feasible path distances from the robot to all key points; Step S2.3 specifically includes: Step S2.3.1: Define a multi-dimensional evaluation index, the weight vector corresponding to each index in the multi-dimensional evaluation index, and the preset threshold for screening; Step S2.3.2: Collect the raw data of each robot in the robot cluster on the multi-dimensional evaluation indicators; Step S2.3.3: Based on the preset threshold, perform anomaly elimination on the robots to obtain a set of valid candidate robots; Step S2.3.4: Calculate the quantitative score for each raw data item corresponding to the multi-dimensional evaluation index for each robot in the set of valid candidate robots; Step S2.3.5: Based on the quantitative scores and weight vectors of each indicator in the multi-dimensional evaluation index, calculate the comprehensive priority score of each robot and sort them according to the comprehensive priority score; Step S2.3.6: Based on the workload of the target area, the working efficiency of a single robot, and the working time requirement, calculate the minimum number of robots required, and determine this number as the final number of robots participating in the operation; In step S2.4, robots are selected sequentially from the set of valid candidate robots according to the priority ranking results from high to low, until the number of selected robots reaches the final number of robots participating in the operation. The selected robots are the task robots.

[0010] Furthermore, step S3 includes: Step S3.1: Obtain one or more feasible paths within the target area; Step S3.2: Perform geometric expansion on the feasible path to form a path coverage area that covers the target area; Step S3.3: Divide the path coverage area into an equal number of sub-regions according to the number of task robots.

[0011] Furthermore, in step S3.1, the feasible path is a robot-accessible path extracted from the global map and marked within the target area, which has avoided the obstacle areas marked in the global map; In step S3.2, the geometric expansion process involves extending a predetermined expansion width to both sides of the centerline of the feasible path as a baseline. Step S3.3 specifically includes: dividing the path coverage area equally based on the total extension length of the path coverage area formed after the feasible path expansion and the number of task robots to obtain an equal number of sub-regions.

[0012] Furthermore, step S4 includes: Step S4.1: Determine a matching reference point for each of the sub-regions; Step S4.2: For each of the task robots, calculate the matching cost for reaching the matching reference point in each sub-region; Step S4.3: Based on the calculated matching costs, assign a task robot to each sub-region to achieve a one-to-one correspondence.

[0013] Furthermore, in step S4.1, the matching reference point is the geometric center point of the corresponding sub-region, which is obtained by calculating the average vertex coordinates of the polygonal region occupied by the sub-region in the global map; In step S4.2, the matching cost is the actual reachable path distance from the task robot's current position to the sub-region matching reference point; the actual reachable path distance is the length of the shortest feasible path planned based on the obstacle area in the global map; Step S4.3 specifically includes: for each sub-region, selecting the robot with the lowest matching cost to reach the matching reference point of the sub-region from all task robots and assigning it; if multiple task robots have the same minimum matching cost, selecting the robot with the shorter straight-line distance to the matching reference point; if the straight-line distance is also the same, selecting the robot with higher remaining energy.

[0014] Furthermore, step S5 includes: Step S5.1: Obtain the job constraints and environmental feature information corresponding to each sub-region; Step S5.2: Based on the aforementioned operational constraints and environmental characteristics, construct a path planning model for each sub-region; Step S5.3: Solve the path planning model to generate the optimal operation route for each task robot in its corresponding sub-region; Step S5.4: Send the global task and the corresponding optimal task route to the corresponding task robot.

[0015] Furthermore, in step S5.1, the operation constraints are obtained by parsing the global operation task, and include at least one of the following: coverage number requirement, restricted area, necessary node, and operation time limit; the environmental feature information is obtained from the global map, and includes at least one of the following: terrain type, obstacle location, and key facility location. In step S5.2, the path planning model includes an objective function, which is configured to comprehensively optimize at least two of the following: the coverage integrity of the path to sub-regions, the total length of the path, and the terrain passability of the path. In step S5.3, the path planning model is solved using a heuristic search algorithm or a mathematical optimization algorithm to obtain the optimal operation route; In step S5.4, task and route data are sent to each task robot via a wireless communication network.

[0016] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are: This invention employs a five-step process: receiving a global task containing a target area, determining the task robot based on the global map and the position and status information of each robot in the robot cluster, dividing the target area into sub-regions matching the number of task robots according to feasible paths, establishing a one-to-one correspondence between sub-regions and task robots based on distance matching rules, and planning the sub-region operation route for each task robot. This process eliminates the fixed partitioning and preset path modes.

[0017] Breaking away from the limitations of fixed work areas and paths in existing technologies, this technology enables dynamic division of target areas, intelligent selection of task robots, optimal matching of sub-areas and task robots, and autonomous planning of work paths. Therefore, when global work tasks change, there is no need for manual replanning, which solves the defects of cumbersome operation, low efficiency and poor adaptability, and improves the flexibility and efficiency of robot swarm operations. Attached Figure Description

[0018] Figure 1 This is a flowchart of the present invention. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0020] Therefore, the following detailed description of the embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the invention without inventive effort are within the scope of protection of the invention.

[0021] The route determination method of the present invention can be applied to scenarios such as sanitation cleaning, park inspection, and warehouse sorting. The present invention will take sanitation cleaning as an example to specifically illustrate the solution of the present invention.

[0022] like Figure 1 As shown, this invention discloses a method for determining the route of a robot swarm, the method comprising: Step S1: Receive a global job task, wherein the global job task contains at least a target area; Step S2: Based on the global map and the position and status information of each robot in the robot cluster, determine multiple task robots to perform the global task; Step S3: Based on the feasible paths within the target area, divide the target area into multiple sub-regions equal to the number of task robots; Step S4: Based on the distance matching rules, map each of the multiple sub-regions to one of the multiple task robots; Step S5: Plan the operation route in the corresponding sub-region for each task robot.

[0023] This invention employs a five-step process: receiving a global task containing a target area, determining the task robot based on the global map and the position and status information of each robot in the robot cluster, dividing the target area into sub-regions matching the number of task robots according to feasible paths, establishing a one-to-one correspondence between sub-regions and task robots based on distance matching rules, and planning the sub-region operation route for each task robot. This process eliminates the fixed partitioning and preset path modes.

[0024] Breaking away from the limitations of fixed work areas and paths in existing technologies, this technology enables dynamic division of target areas, intelligent selection of task robots, optimal matching of sub-areas and task robots, and autonomous planning of work paths. Therefore, when global work tasks change, there is no need for manual replanning, which solves the defects of cumbersome operation, low efficiency and poor adaptability, and improves the flexibility and efficiency of robot swarm operations.

[0025] Furthermore, step S1 specifically includes: Step S1.1: Receive task configuration information from the user client; Step S1.2: Parse the task configuration information to generate structured global task data, wherein the global task data includes at least the range information of the target area; Step S1.3: Verify the integrity and validity of the global task data. If the verification passes, store the global task data and generate a global task based on the verified global task data for subsequent steps.

[0026] This invention ensures that the input global task data is accurate, effective, and structured by parsing and verifying the task configuration information. This provides a unified and reliable data input foundation for subsequent steps such as task robot selection and sub-region division, avoiding errors in subsequent processes due to data chaos.

[0027] Furthermore, in step S1.1, the user client is a mobile terminal APP or a computer application; the task configuration information is generated by the user through at least one of the following methods: drawing an area on an electronic map, selecting key points, or inputting parameters.

[0028] This invention supports users in configuring tasks through multiple terminals and methods, adapting to different users' operating habits and usage scenarios, improving the convenience and flexibility of global task configuration, and reducing the operational threshold for task issuance.

[0029] Furthermore, in step S1.2, the parsed global task data also includes at least one of the following: task area range, number of tasks, expected task duration, task type, expected number of robots to participate, and designated robot number.

[0030] To enrich the dimensions of global task data, supplement it with at least one of the following: task area range, number of tasks, expected task duration, task type, expected number of robots involved, and designated robot ID. This provides a more comprehensive basis for task constraints in subsequent steps. For example, the number of tasks can guide the coverage of path planning, and the expected task duration can help determine the number of robots required for the task, making the entire process of robot cluster operations more closely aligned with actual task requirements.

[0031] Furthermore, the operating area is larger than the target area and includes resource replenishment points, which include at least one of waste collection points and charging points.

[0032] The defined operational area is larger than the target area and includes at least one resource replenishment point, such as a waste collection point or a charging point, thus expanding the functional boundaries of the operational area. This provides resource replenishment support for the task robots during operation, preventing them from interrupting their work due to insufficient power or full waste load, and ensuring the continuity and stability of robot swarm operations.

[0033] Furthermore, the types of operations include at least one of litter collection and area patrol.

[0034] Furthermore, step S2 specifically includes: Step S2.1: Obtain the global map and extract the target area, the key points contained in the target area, and the position and status information of each robot in the robot cluster from it; Step S2.2: For each robot in the robot cluster, based on its location information, the key points, and the obstacle areas marked in the global map, plan a feasible path and calculate the average reachable distance from the robot to each key point in the target area; Step S2.3: Prioritize each robot in the robot cluster based on multi-dimensional evaluation indicators to obtain priority ranking results. The multi-dimensional evaluation indicators include at least average reachability distance and status information. Combined with the workload of the target area, the robot's work efficiency and the work time requirements of the global task, determine the final number of robots participating in the operation. Step S2.4: Based on the determined number of robots and priority ranking results, select a corresponding number of robots from the robot cluster as the task robots.

[0035] This invention establishes a systematic task robot screening mechanism. By comprehensively analyzing the status and reachability of each robot in the robot cluster, it accurately selects task robots suitable for operation in the target area, ensuring that the selected task robots have the ability to complete the task. At the same time, by determining a reasonable number of task robots, it avoids robot resource redundancy or insufficiency.

[0036] Furthermore, in step S2.1: The global map is marked with obstacle areas and passable areas; The key points are the necessary passageways for the robot to enter the target area or the core operation nodes within the target area; The location information is the real-time coordinates of each robot, and the status information includes at least one of the following: the remaining energy of each robot, the current task status, and the path terrain adaptation information.

[0037] The above process provides accurate and detailed basic data for feasible path planning and robot prioritization. For example, obstacle areas can guide the generation of obstacle avoidance paths, and remaining energy can reflect the endurance of the task robot, ensuring the scientific and accurate nature of the subsequent task robot selection process.

[0038] Furthermore, in step S2.2: The feasible path is a path planned by the robot from its current position to key points in the target area, with the obstacle area as a constraint. The average reachable distance is the arithmetic mean of the feasible path distances from the robot to all key points.

[0039] Furthermore, sub-step S2.3 specifically includes: Step S2.3.1: Define a multi-dimensional evaluation index, the weight vector corresponding to each index in the multi-dimensional evaluation index, and the preset threshold for screening; Step S2.3.2: Collect the raw data of each robot in the robot cluster on the multi-dimensional evaluation indicators; Step S2.3.3: Based on the preset threshold, perform anomaly elimination on the robots to obtain a set of valid candidate robots; Step S2.3.4: Calculate the quantitative score for each raw data item corresponding to the multi-dimensional evaluation index for each robot in the set of valid candidate robots; Step S2.3.5: Based on the quantitative scores and weight vectors of each indicator in the multi-dimensional evaluation index, calculate the comprehensive priority score of each robot and sort them according to the comprehensive priority score; Step S2.3.6: Based on the workload of the target area, the working efficiency of a single robot, and the working time requirement, calculate the minimum number of robots required, and determine this number as the final number of robots participating in the operation.

[0040] This invention enables the scientific screening and quantitative evaluation of robots in a robot swarm through a standardized process, accurately determining the minimum number of task robots required to meet the operation time requirements, thus solving the problem of existing technologies being unable to quickly and reasonably screen task robots and determine the number of task robots.

[0041] Furthermore, in step S2.4, robots are selected sequentially from the set of valid candidate robots according to the priority ranking results from high to low, until the number of selected robots reaches the final number of robots participating in the operation. The selected robots are the task robots.

[0042] This invention ensures that the selected task robots are the optimal combination from the effective candidate robot set, maximizes the overall operational capability of the robot cluster, and guarantees the efficient advancement of global tasks.

[0043] Furthermore, the multi-dimensional evaluation metrics include average reachability, remaining energy, path environment complexity, and current task status; The weight vector is W=[W1,W2,W3,W4], where W1, W2, W3, and W4 are the weight coefficients corresponding to the average reachability distance, remaining energy, path environment complexity, and current task state, respectively, and satisfy W1+W2+W3+W4=1.

[0044] This invention comprehensively evaluates robots from four dimensions: accessibility, battery life, path adaptability, and task occupancy. This avoids the one-sidedness of evaluation based on a single indicator. By allocating weights, the importance of each indicator can be flexibly adjusted to adapt to the evaluation needs of different work scenarios.

[0045] Furthermore, the preset thresholds include an energy threshold B_threshold and a task time conflict threshold T_conflict; The default value for the energy threshold B_threshold is 20%. The task time conflict threshold T_conflict is equal to the task time T_task specified by the global task.

[0046] By excluding robots with insufficient power using an energy threshold and excluding robots whose current task conflicts with the global task using a task time conflict threshold, invalid robots are effectively filtered out, ensuring that the robots in the set of valid candidate robots are feasible for operation.

[0047] Furthermore, the formula for calculating the quantization score S_Di of the average reachable distance is: When D_max ≠ D_min, S_Di = (D_max - Di) / (D_max - D_min); When D_max = D_min, S_Di = 1; Where D_i is the average reachable distance from robot i to each key point in the target area, D_max is the maximum value of D_i among all valid candidate robots, and D_min is the minimum value of D_i among all valid candidate robots.

[0048] This invention employs an extreme value normalization method to calculate the average reachable distance quantification score S_Di. When D_max ≠ D_min, the score is calculated as (D_max - Di) / (D_max - D_min), with shorter distances receiving higher scores. When D_max = D_min, the score is uniformly set to 1, thus achieving standardized quantification of the average reachable distance index. This transforms the average reachable distances of different robots into comparable scores within the 0-1 range, eliminating dimensional differences in distance values ​​between different robots and providing a unified quantification standard for comprehensive priority scoring.

[0049] Furthermore, the formula for calculating the quantification score S_Ei of the remaining energy is: S_Ei=B_i / 100; Where B_i is the current remaining energy percentage of robot i, and 0≤B_i≤100.

[0050] This invention sets the remaining energy quantification score S_Ei as the ratio of the remaining energy percentage B_i to 100, directly reflecting the battery endurance as a percentage of battery power. This provides an intuitive and simple way to quantify the robot's battery endurance; the higher the remaining battery power, the higher the score. This allows the battery endurance indicator to directly participate in the comprehensive priority scoring, improving the intuitiveness of the evaluation results.

[0051] Furthermore, the calculation process for the quantification score S_Ci of the path environment complexity is as follows: (1) Calculate the average environmental resistance coefficient R_i, the formula is: R_i=Σ(l_ik×r_k) / Σl_ik; Where l_ik is the length of the k-th segment on the feasible path of robot i, r_k is the preset resistance coefficient of the environment type corresponding to the segment, and Σ represents the summation operation; (2) Calculate the path environment complexity score S_Ci, using the following formula: When R_max ≠ R_min, S_Ci = (R_max - R_i) / (R_max - R_min); When R_max = R_min, S_Ci = 1; Where R_max is the maximum value of R_i among all valid candidate robots, and R_min is the minimum value of R_i among all valid candidate robots.

[0052] This invention calculates the average environmental resistance coefficient R_i and then uses an extreme value normalization method to calculate the quantitative score S_Ci of path environmental complexity. The smaller the resistance coefficient, the higher the score, thus establishing a quantitative evaluation process for path environmental complexity. This invention transforms the impact of path terrain on robots into comparable quantitative scores, accurately reflecting the traversal difficulty of different robots on corresponding paths. This makes the evaluation of robot adaptability to the path environment more scientific and improves the accuracy of task robot selection.

[0053] Furthermore, the correspondence between the preset resistance coefficient r_k and the terrain type is as follows: 1 for flat roads, 1.5 for gravel roads or dirt roads, 2 for steps, 2.5 for steep slopes, and 3 for shallow water.

[0054] This invention establishes a standard for the correspondence between terrain types and resistance coefficients by clearly defining the preset resistance coefficient r_k values ​​corresponding to terrain types such as flat roads, gravel / dirt roads, steps, steep slopes, and shallow water accumulation. This provides a clear and unified basis for calculating the average environmental resistance coefficient R_i, avoiding distortion in path environmental complexity scoring caused by inconsistent resistance coefficient values, and ensuring the objectivity and consistency of the evaluation results.

[0055] Furthermore, the calculation rule for the quantization score S_Ti of the current task state is as follows: (1) If robot i is in an idle state, then S_Ti=1; (2) If robot i is in a busy state, then S_Ti = 1 - t_rem_i / t_total_i; Where t_rem_i is the remaining time of robot i's current task, and t_total_i is the total time of robot i's current task.

[0056] This invention quantifies the robot's task occupancy by setting a calculation rule for the current task status score S_Ti. The idle state scores 1, while the busy state scores according to 1 - t_rem_i / t_total_i. The higher the percentage of remaining task time, the lower the score, thus quantifying the robot's task occupancy. This accurately distinguishes the job availability of idle and busy robots. The score of a busy robot decreases as the remaining task time increases, effectively filtering out robots with low task occupancy that can be quickly deployed to new tasks, ensuring the timely progress of overall tasks.

[0057] Furthermore, the formula for calculating the comprehensive priority score S_i is as follows: S_i=W1×S_Di+W2×S_Ei+W3×S_Ci+W4×S_Ti; Where S_Di, S_Ei, S_Ci, and S_Ti are the average reachable distance, remaining energy, path environment complexity, and current task state quantification scores of robot i, respectively. The priority sorting rules are as follows: sort by S_i from high to low; if S_i are the same, sort by D_i from small to large; if D_i are the same, sort by B_i from large to small.

[0058] This invention calculates the comprehensive priority score using the weighted summation formula S_i=W1×S_Di+W2×S_Ei+W3×S_Ci+W4×S_Ti, and sets a sorting rule in the case of the same score: from smallest to largest average reachable distance and from largest to smallest remaining energy, thus establishing a precise robot priority sorting mechanism.

[0059] A scientific and orderly robot priority sequence is generated, and the tie-breaking rules further refine the priority differences of robots, providing a clear and unambiguous basis for the selection of task robots and ensuring that the selected task robot combination is optimal.

[0060] Furthermore, in step S2.3.6, the formula for calculating the minimum number of robots N_min is: N_min=ceil(S_target / (V_single×T_task)); Where S_target is the workload of the target area, V_single is the work efficiency of a single robot, T_task is the work time requirement, and ceil() is the round-up function; The final number of robots participating in the operation is the smaller value between N_min and the number of valid candidate robots in the robot cluster, and is not less than 1.

[0061] This invention accurately calculates the minimum number of task robots required to meet the operation time requirements, avoiding resource waste due to an excessive number of task robots or an insufficient number that prevents the completion of the global operation task on time. At the same time, it combines the adjustment of the number of effective candidate robots to ensure the feasibility of the determined number.

[0062] Furthermore, step S3 specifically includes: Step S3.1: Obtain one or more feasible paths within the target area; Step S3.2: Perform geometric expansion on the feasible path to form a path coverage area that covers the target area; Step S3.3: Divide the path coverage area into an equal number of sub-regions according to the number of task robots.

[0063] This invention covers the target area based on the geometric expansion of feasible paths, and the number of sub-regions is perfectly matched with the number of task robots. This breaks the limitation of fixed sub-region boundaries, realizes dynamic partitioning of the target area, and makes the working range of each task robot balanced and reasonable.

[0064] Furthermore, in step S3.1, the feasible path is a robot-accessible path extracted from the global map and marked within the target area, which has avoided the obstacle areas marked in the global map.

[0065] Furthermore, the geometric expansion process in step S3.2 involves extending a preset expansion width d to both sides of the centerline of the feasible path as a baseline, where d is a positive number set according to the robot's physical dimensions or working width.

[0066] This invention matches the expanded path coverage area with the robot's actual working width, ensuring that the task robot can achieve full coverage when working in a sub-area, and avoiding omissions or repetitions in the work due to the coverage area being too wide or too narrow.

[0067] Furthermore, step S3.3 specifically involves dividing the path coverage area into equal parts based on the total extension length of the path coverage area formed after the feasible path expansion and the number of task robots, thereby obtaining an equal number of sub-regions.

[0068] This invention ensures that the extension length of each sub-region is balanced after division, thus ensuring that the workload of each task robot is basically the same and avoiding the problem of uneven workload where some task robots have too much workload and some task robots have too little workload.

[0069] Furthermore, step S3.2 specifically includes: Step S3.2.1: Discretize the centerline of the feasible path into a polyline defined by an ordered sequence of key points; Step S3.2.2: For each straight line segment on the fold line, generate a rectangular expansion area with a width of 2d centered on that straight line segment; Step S3.2.3: For the corners between adjacent straight line segments, generate a fan-shaped expansion area with radius d to fill the gaps between them; Step S3.2.4: Merge all rectangular expansion regions and sector expansion regions to form the path coverage area.

[0070] Step S3.2 includes a four-step process: using the centerline of the discrete feasible path as an ordered key point polyline, generating rectangular expansion regions for straight line segments, generating fan-shaped expansion regions at corners, and merging them to form the path coverage area. This establishes a refined path expansion mechanism. By covering straight line segments with rectangular expansion regions and filling corner gaps with fan-shaped expansion regions, gapless construction of the path coverage area is achieved, ensuring that the target area is completely covered and improving the accuracy of sub-region division.

[0071] Furthermore, in step S3.2.1, the ordered key point sequence is P={p1,p2,...,pn}, where each key point p_i=(x_i,y_i) is a planar coordinate, and adjacent key points p_i and p_{i+1} form a straight line segment L_i, where i is the key point index.

[0072] This step clarifies that the ordered keypoint sequence P={p1,p2,...,pn} is composed of planar coordinate keypoints, with adjacent keypoints forming a straight line segment Li, refining the structure of the polyline. This provides a precise geometric coordinate basis for the generation of rectangular and sector-shaped expansion regions, ensuring that the generated location of the expansion region perfectly matches the feasible path, thus improving the geometric accuracy of the path-covered area.

[0073] Furthermore, for a line segment L_i of length |L_i|, the area S_rect_i of the rectangular expansion region it generates is calculated as follows: S_rect_i = |L_i| × 2d, where |L_i| is the length of the line segment L_i.

[0074] This invention calculates the area of ​​the rectangular expansion region of a straight line segment using the formula S_rect_i=|Li|×2d, quantifying the area parameter of the rectangular expansion region. This provides accurate rectangular area data for calculating the total area of ​​the path coverage area, making the load assessment of subsequent sub-region division more quantitatively based and improving the scientific rigor of sub-region division.

[0075] Furthermore, for a corner with an included angle of θ_i (in radians), the area S_sector_i of the resulting sector expansion region is calculated as follows: S_sector_i = 0.5 × d² × θ_i, where d² represents the square of the expansion width d.

[0076] This invention calculates the area of ​​the corner sector expansion zone using the formula S_sector_i=0.5×d²×θ_i, quantifying the area parameter of the sector expansion zone. It accurately calculates the area of ​​the sector expansion zone at the corner, filling the gaps in the rectangular expansion zone at the corner, and simultaneously provides complete data for calculating the total area of ​​the path coverage area, ensuring the accuracy of the area calculation.

[0077] Furthermore, the included angle θ_i is calculated as follows: calculate the unit direction vector u_i of line segment L_i and the unit direction vector u_{i+1} of line segment L_{i+1}; then θ_i = arccos(u_i·u_{i+1}).

[0078] This invention calculates the corner angle θ_i by taking the inverse cosine of the dot product of the unit direction vectors of adjacent straight line segments, thus establishing a method for calculating the angle. It precisely quantifies the angle parameters of feasible path corners, providing accurate angle input for calculating the area of ​​the sector expansion zone, ensuring a perfect match between the sector expansion zone and the corner shape, and improving the fit of the path coverage area.

[0079] Furthermore, the total area S_total of the path-covered region is calculated using the following formula: S_total = ΣS_rect_i + ΣS_sector_j, where ΣS_rect_i represents the summation of the areas of all rectangular expansion regions, and ΣS_sector_j represents the summation of the areas of all sector expansion regions.

[0080] This invention calculates the total area of ​​the path-covered region using the formula S_total=ΣS_rect_i+ΣS_sector_j, thus summing the areas of the rectangular and sector-shaped expansion zones. This precise quantification of the overall operational area of ​​the target region provides a quantitative basis for load balancing assessment during sub-region division, ensuring that the operational areas of each sub-region are essentially consistent and improving the load balancing of the task robot's operations.

[0081] Furthermore, step S4 specifically includes: Step S4.1: Determine a matching reference point for each of the sub-regions; Step S4.2: For each of the task robots, calculate the matching cost for reaching the matching reference point in each sub-region; Step S4.3: Based on the calculated matching costs, assign a task robot to each sub-region to achieve a one-to-one correspondence.

[0082] This invention quantifies the movement cost of a task robot to a sub-region by matching cost, achieving optimal matching between the sub-region and the task robot, reducing the ineffective movement loss of the task robot traveling back and forth between different sub-regions, and improving the efficiency of robot swarm operations.

[0083] Furthermore, in step S4.1, the matching reference point is the geometric center point of the corresponding sub-region, which is obtained by calculating the average vertex coordinates of the polygonal region occupied by the sub-region in the global map.

[0084] The matching reference point is defined as the average coordinate of the vertices of the polygon in the sub-region, i.e., the geometric center point of the sub-region, and the selection rules for the reference point are established. This ensures that the matching reference point represents the core location of the sub-region, guarantees that the calculated matching cost is representative, and avoids distortion of the matching results due to unreasonable reference point selection.

[0085] Furthermore, in step S4.2, the matching cost is the actual reachable path distance from the task robot's current position to the sub-region matching reference point; the actual reachable path distance is the length of the shortest feasible path planned based on the obstacle area in the global map.

[0086] This invention uses the shortest feasible path distance as the matching basis, objectively reflects the difficulty of the task robot moving to the sub-region, provides accurate quantitative indicators for task robot-sub-region matching, and improves the optimality of the matching scheme.

[0087] Furthermore, step S4.3 specifically involves: for each sub-region, selecting the robot with the lowest matching cost from all task robots to reach the matching reference point of that sub-region for allocation; if multiple task robots have the same minimum matching cost, selecting the robot with the shorter straight-line distance to the matching reference point; if the straight-line distance is also the same, selecting the robot with higher remaining energy.

[0088] Task robots are assigned based on the principle of minimizing matching cost. When costs are the same, they are sorted according to the rule of shorter straight-line distance and higher remaining energy, thus establishing the priority of task robot allocation. This ensures that each sub-region is matched with the task robot with the lowest movement cost. Tie-breaking rules further optimize the matching results, minimizing the overall movement loss of task robots and improving the overall efficiency of robot swarm operations.

[0089] Furthermore, step S5 specifically includes: Step S5.1: Obtain the job constraints and environmental feature information corresponding to each sub-region; Step S5.2: Based on the aforementioned operational constraints and environmental characteristics, construct a path planning model for each sub-region; Step S5.3: Solve the path planning model to generate the optimal operation route for each task robot in its corresponding sub-region; Step S5.4: Send the global task and the corresponding optimal task route to the corresponding task robot.

[0090] This invention customizes the optimal operation route based on the specific constraints and environmental characteristics of each sub-region, realizes autonomous planning of the operation path, solves the defect of fixed path in the prior art, and improves the operation efficiency of the task robot in the sub-region.

[0091] Furthermore, in step S5.1: The job constraints are obtained by parsing the global job task and include at least one of the following: coverage requirement, restricted area, required node, and job time limit; The environmental feature information is obtained from the global map and includes at least one of the following: terrain type, obstacle location, and key facility location.

[0092] The above method provides comprehensive constraints for the path planning model, ensuring that the generated operation route meets the requirements of the overall operation task and environmental characteristics, avoiding the route from violating restrictions such as restricted areas and operation time limits, and improving the compliance and feasibility of the route.

[0093] Furthermore, in step S5.2, the path planning model includes an objective function configured to comprehensively optimize at least two of the following optimization objectives: the completeness of path coverage of sub-regions, the total length of the path, and the terrain passability of the path.

[0094] This invention constructs a multi-objective optimization path planning system by comprehensively optimizing at least two of the following: the path's coverage integrity to sub-regions, the total path length, and the path's terrain traversability, through a defined objective function of the path planning model. This avoids the limitations of single-objective optimization, such as balancing coverage integrity and path length, and can shorten the operational path while ensuring full operational coverage, generating an optimal operational route that better meets actual needs.

[0095] Furthermore, the expression for the objective function F is: F = α × f_cov + β × f_len + γ × f_ter, where: f_cov is a function that characterizes the integrity of path coverage, and its value is negatively correlated with the coverage rate. f_len is the total path length; f_ter is a function that characterizes terrain passability, and its value comprehensively reflects the difficulty of traversing the terrain traversed by the path; α, β, and γ are the weighting coefficients corresponding to f_cov, f_len, and f_ter, respectively.

[0096] This invention constructs the objective function using the weighted summation formula F=α×f_cov+β×f_len+γ×f_ter, and adjusts the priority of each optimization objective through weight coefficients α, β, and γ, establishing a quantitative calculation method for multi-objective optimization. This allows the path planning model to adjust the priority of each objective according to the needs of different operational scenarios. For example, in sanitation cleaning scenarios, the weight of coverage integrity can be increased, while in park inspection scenarios, the weight of path length can be increased, improving the model's scenario adaptability.

[0097] Furthermore, in step S5.3, the path planning model is solved using a heuristic search algorithm or a mathematical optimization algorithm to obtain the optimal operation route.

[0098] The heuristic search algorithm includes at least one of the following: A algorithm, improved Dijkstra algorithm, ant colony algorithm, particle swarm optimization algorithm, RRT algorithm, and PRM algorithm.

[0099] The mathematical optimization algorithm includes at least one of linear programming algorithm, integer linear programming algorithm, convex optimization algorithm, mixed integer programming algorithm, and genetic algorithm.

[0100] This invention establishes a solution method for the path planning model by using heuristic search algorithms or mathematical optimization algorithms. The algorithm efficiently solves multi-objective optimization problems, quickly generating optimal operation routes for sub-regions, ensuring the timeliness of path planning, and meeting the time requirements of the overall task.

[0101] Furthermore, in step S5.4, the task and route data are sent to each task robot via a wireless communication network.

[0102] The global task and optimal route are distributed to the task robot via a wireless communication network, establishing a method for distributing global tasks and routes. This enables remote and rapid distribution of global tasks and routes without the need for manual on-site configuration, improving the convenience and timeliness of global task execution and ensuring the efficient and collaborative advancement of robot swarm operations.

[0103] Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for determining the route of a robot swarm, characterized in that, The method includes: Step S1: Receive a global job task, wherein the global job task contains at least a target area; Step S2: Based on the global map and the position and status information of each robot in the robot cluster, determine multiple task robots to perform the global task; Step S3: Based on the feasible paths within the target area, divide the target area into multiple sub-regions equal to the number of task robots; Step S4: Based on the distance matching rules, map each of the multiple sub-regions to one of the multiple task robots; Step S5: Plan the operation route in the corresponding sub-region for each task robot.

2. The method for determining the route of a robot swarm according to claim 1, characterized in that, Step S1 specifically includes: Step S1.1: Receive task configuration information from the user client; Step S1.2: Parse the task configuration information to generate structured global task data, wherein the global task data includes at least the range information of the target area; Step S1.3: Verify the integrity and validity of the global task data. If the verification passes, store the global task data and generate a global task based on the verified global task data.

3. The method according to claim 1, characterized in that, Step S2 includes: Step S2.1: Obtain the global map and extract the target area, the key points contained in the target area, and the position and status information of each robot in the robot cluster from it; Step S2.2: For each robot in the robot cluster, based on its location information, the key points, and the obstacle areas marked in the global map, plan a feasible path and calculate the average reachable distance from the robot to each key point in the target area; Step S2.3: Prioritize each robot in the robot cluster based on multi-dimensional evaluation indicators to obtain priority ranking results. The multi-dimensional evaluation indicators include at least average reachability distance and status information. Combined with the workload of the target area, the robot's work efficiency and the work time requirements of the global task, determine the final number of robots participating in the operation. Step S2.4: Based on the determined number of robots and priority ranking results, select a corresponding number of robots from the robot cluster as the task robots.

4. The method according to claim 3, characterized in that, In step S2.1, the global map is marked with obstacle areas and passable areas; the key points are the necessary passage points for the robot to enter the target area or the core operation nodes within the target area; the location information is the real-time coordinates of each robot; and the status information includes at least one of the following: the remaining energy of each robot, the current task status, and the path terrain adaptation information. In step S2.2, the feasible path is the obstacle avoidance and detour path planned by the robot from its current position to each key point in the target area, with the obstacle area as the constraint condition; the average reachable distance is the arithmetic mean of the feasible path distances from the robot to all key points; Step S2.3 specifically includes: Step S2.3.1: Define a multi-dimensional evaluation index, the weight vector corresponding to each index in the multi-dimensional evaluation index, and the preset threshold for screening; Step S2.3.2: Collect the raw data of each robot in the robot cluster on the multi-dimensional evaluation indicators; Step S2.3.3: Based on the preset threshold, perform anomaly elimination on the robots to obtain a set of valid candidate robots; Step S2.3.4: Calculate the quantitative score for each raw data item corresponding to the multi-dimensional evaluation index for each robot in the set of valid candidate robots; Step S2.3.5: Based on the quantitative scores and weight vectors of each indicator in the multi-dimensional evaluation index, calculate the comprehensive priority score of each robot and sort them according to the comprehensive priority score; Step S2.3.6: Based on the workload of the target area, the working efficiency of a single robot, and the working time requirement, calculate the minimum number of robots required, and determine this number as the final number of robots participating in the operation; In step S2.4, robots are selected sequentially from the set of valid candidate robots according to the priority ranking results from high to low, until the number of selected robots reaches the final number of robots participating in the operation. The selected robots are the task robots.

5. The method according to claim 1, characterized in that, Step S3 includes: Step S3.1: Obtain one or more feasible paths within the target area; Step S3.2: Perform geometric expansion on the feasible path to form a path coverage area that covers the target area; Step S3.3: Divide the path coverage area into an equal number of sub-regions according to the number of task robots.

6. The method for determining the route of a robot swarm according to claim 5, characterized in that, In step S3.1, the feasible path is a robot-accessible path extracted from the global map and marked within the target area, which has avoided the obstacle areas marked in the global map; In step S3.2, the geometric expansion process involves extending a predetermined expansion width to both sides of the centerline of the feasible path as a baseline. Step S3.3 specifically includes: dividing the path coverage area equally based on the total extension length of the path coverage area formed after the feasible path expansion and the number of task robots to obtain an equal number of sub-regions.

7. The method according to claim 1, characterized in that, Step S4 includes: Step S4.1: Determine a matching reference point for each of the sub-regions; Step S4.2: For each of the task robots, calculate the matching cost for reaching the matching reference point in each sub-region; Step S4.3: Based on the calculated matching costs, assign a task robot to each sub-region to achieve a one-to-one correspondence.

8. The method according to claim 7, characterized in that, In step S4.1, the matching reference point is the geometric center point of the corresponding sub-region, which is obtained by calculating the average vertex coordinates of the polygonal region occupied by the sub-region in the global map; In step S4.2, the matching cost is the actual reachable path distance from the task robot's current position to the sub-region matching reference point; the actual reachable path distance is the length of the shortest feasible path planned based on the obstacle area in the global map; Step S4.3 specifically includes: for each sub-region, selecting the robot with the lowest matching cost to reach the matching reference point of the sub-region from all task robots and assigning it; if multiple task robots have the same minimum matching cost, selecting the robot with the shorter straight-line distance to the matching reference point; if the straight-line distance is also the same, selecting the robot with higher remaining energy.

9. The method according to claim 1, characterized in that, Step S5 includes: Step S5.1: Obtain the job constraints and environmental feature information corresponding to each sub-region; Step S5.2: Based on the aforementioned operational constraints and environmental characteristics, construct a path planning model for each sub-region; Step S5.3: Solve the path planning model to generate the optimal operation route for each task robot in its corresponding sub-region; Step S5.4: Send the global task and the corresponding optimal task route to the corresponding task robot.

10. The method according to claim 9, characterized in that, In step S5.1, the operation constraints are obtained by parsing the global operation task, and include at least one of the following: coverage number requirement, restricted area, necessary node, and operation time limit; the environmental feature information is obtained from the global map, and includes at least one of the following: terrain type, obstacle location, and key facility location. In step S5.2, the path planning model includes an objective function, which is configured to comprehensively optimize at least two of the following: the coverage integrity of the path to sub-regions, the total length of the path, and the terrain passability of the path. In step S5.3, the path planning model is solved using a heuristic search algorithm or a mathematical optimization algorithm to obtain the optimal operation route; In step S5.4, task and route data are sent to each task robot via a wireless communication network.

Images