Active control method of special robot machining system combined with game optimization strategy

Abstract

The present application relates to the technical field of robot control, in particular to an active control method of special robot processing system combined with game optimization strategy, and the specific steps are as follows: matching robots corresponding to each task to form corresponding matching pairs to construct a feasible matching list; constructing robot list A, assigning the task of the only robot in the constructed robot list A to a strategy set, updating the robot list A to robot list B; screening the feasible matching set with the least number of robots in the robot list B, defining the corresponding feasible matching set as robot list C; using branch iteration assignment method to assign corresponding robots to the tasks of the feasible matching set in the robot list C; using double income function to calculate the adaptation score of each robot corresponding to the task in the assignment process, then calculating the total score through the adaptation score, and assigning through the total score; determining the order of the feasible matching set in the robot list C before the branch iteration assignment method is assigned through the adaptation score.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of robot control, in particular to an active control method of a special robot processing system combined with a game optimization strategy. BACKGROUND

[0002] Special robots are widely used in aerospace precision machining and other extreme scenarios. For example, aerospace material printing requires processing high-strength difficult-to-machine materials and meeting micron-level precision. At the same time, the large-scale and complex structure of aerospace components is obvious, and a single robot is difficult to adapt to the time limit, so the task needs to be executed by multiple robots in collaboration (such as large fuel tank printing, which is divided into forming, welding, detection and other sub-tasks by corresponding special robots).

[0003] Because the aerospace task is related to the core performance of the spacecraft, the robot processing precision (adapted to the micron-level tolerance of the key component) and the maximum processing range (covering large components) are required to be extremely high, and most robots have been pre-calibrated to meet the parameter threshold, which is easy to adapt to multiple robots for the same task. In the case of multiple robots that can adapt to the same task, the multi-robot task allocation will face significant challenges in dealing with the multi-task-multi-robot matching problem, specifically:

[0004] In the process of matching robots with corresponding tasks, the requirements of aerospace manufacturing for task execution are not limited to "basic adaptability", but also need to achieve integrated standards of processing quality, energy consumption cost and time efficiency, and different robots have inherent differences in performance in core dimensions such as processing efficiency (such as processing amount per unit time, task completion cycle), energy consumption cost (such as energy consumption coefficient per unit time, long-term running energy consumption accumulation), precision stability (such as micron-level tolerance control ability, processing deviation fluctuation range) and other core dimensions. And there is usually a coupling constraint relationship between multiple core dimensions (such as increasing processing efficiency may lead to a sharp increase in energy consumption, and pursuing extreme precision may prolong the time limit).

[0005] In order to realize the accurate matching of each robot and the corresponding task, the existing solution often relies on a large number of enumeration to verify different task-robot matching strategy combinations one by one, which is specifically manifested as:

[0006] First, based on the robot machining precision threshold, maximum machining range and other basic parameters, all robots with task adaptation ability are selected as a candidate set. Then, different tasks and robots in the candidate set are combined and paired in an enumeration manner. During the pairing process, the processing efficiency (such as the completion time calculated based on the robot running rate and task complexity), energy consumption cost (such as the product of the unit energy consumption coefficient and the processing time), precision stability (such as the matching degree of the robot historical machining deviation data and the task tolerance requirement) and other core indicators are calculated for each strategy combination (task-robot). If the calculation results of the indicators of a combination meet the constraint requirements, the combination is included in the feasible strategy library. After all combinations are enumerated and verified, the optimal scheme with the best comprehensive performance is selected from the feasible strategy library.

[0007] However, in the enumeration process, in the current aerospace manufacturing scenario, on the one hand, the tasks to be allocated present multiple types of characteristics (such as tasks covering rocket body cabin forming, satellite support lattice processing, fuel tank flange welding, component defect detection and other different process requirements), and each type of task needs to be matched with a robot with corresponding process capability. On the other hand, there are many robots with basic adaptation ability (because most special robots have completed parameter calibration to meet the core requirements of micron-level precision and large-size operation range, and one task often corresponds to 5-8 candidate robots). The combination of the two directly leads to an exponential growth in the number of strategy combinations of “task-robot” — for example, 12 different process tasks and 10 robots with multi-process adaptation ability, only the basic one-to-one matching combination reaches 120, and if considering that some tasks need to be assisted by multiple robots (such as large component processing needs 1 main processing robot + 2 auxiliary positioning robots), the number of strategy combinations will exceed 1,000.

[0008] Under such a large combination base, enumeration needs to verify multiple dimensions for each strategy combination. Since each strategy combination (task-robot) needs to independently calculate the core indicators such as processing efficiency, energy consumption cost and precision stability, the overall calculation amount will far exceed the system carrying efficiency, not only occupying a large amount of computing resources (such as industrial computers need to operate continuously for 8-12 hours to complete all combination verification), but also greatly prolonging the task allocation period (far exceeding the 24-hour task allocation period requirement of aerospace tasks), ultimately resulting in low allocation efficiency, and even missing the best execution window period of the task due to the long enumeration time;

[0009] Therefore, we propose an active control method for special robot machining systems combined with game optimization strategies. SUMMARY

[0010] The present application aims to solve the key problems of exponential growth of combination number, large enumeration calculation amount, and overage distribution period of existing task-robot matching scheme in multi-robot collaborative processing of aerospace component additive manufacturing, and difficult optimal matching of multi-objective coupling scene and quality problem of components caused by multi-robot same task.

[0011] To achieve the above-mentioned purpose, the active control method of the special robot processing system combined with the game optimization strategy has the following working steps:

[0012] S1, judging whether the robot has the strategy feasibility of selecting the task, thereby matching the robot corresponding to each task to form a corresponding matching pair, and jointly constructing a feasible pairing list from multiple matching pairs;

[0013] S2, receiving the feasible pairing list, screening all robots corresponding to each task, then constructing a robot list A, assigning the task of the unique robot in the robot list A to the strategy set, and updating the robot list A to a robot list B;

[0014] Screening the feasible matching set with the least number of robots in the robot list B, defining the corresponding feasible matching set as a robot list C; using the branch iteration distribution method to assign the corresponding robot to the task of the feasible matching set in the robot list C: forcing the assignment of the robot in the feasible matching set , and using a double-income function to calculate the adaptation score of each robot corresponding to the task, then selecting the robot with high total score through the total score calculated by the adaptation score to execute the corresponding task of the feasible matching set , and eliminating the assigned robots and tasks in other feasible matching sets, and then updating the robot list C again until all tasks in the strategy set are assigned.

[0015] S3, obtaining the robot list C, calling out the robot corresponding to the task in each feasible matching set, and screening the corresponding task in the robot list with the robot as the retrieval condition; if the screened task has calculated the adaptation score through step S2, mark the corresponding task; count the total number of tasks corresponding to all robots and the number of marked tasks, calculate the matching proportion of each feasible matching set, and call out the matching proportion of all feasible matching sets in turn, and select the feasible matching set with the largest matching proportion to assign the robot to the corresponding task through step S2.

[0016] As a further improvement of the technical solution, the step S1 obtains the robot set divided according to different additive processes, and the task package set formed after the aerospace component is disassembled; respectively call out the core parameters of each robot in the robot set For each task, the robot and task are paired using core parameters and task parameters to form matching pairs. Then, multiple matching pairs are used to construct a feasible pairing list. .

[0017] As a further improvement to this technical solution, step S1 constructs matching pairs. Receive task Precision requirements Space requirements and robots Machining accuracy and maximum processing range ; Judge the robot Does it have the ability to select tasks? Feasibility of the strategy:

[0018] If the robot Machining accuracy ≥Task package Precision requirements ,robot Maximum processing range ≥Task package Space requirements When, then determine the robot With task selection The feasibility of the strategy, and for robots Assign tasks Conversely, it determines the robot. It does not have the ability to select tasks. The feasibility of the strategy.

[0019] As a further improvement to this technical solution, step S2 receives the feasible pairing list containing multiple matching pairs from step S1. ;

[0020] For each task, select robots capable of performing the task to form a feasible matching set for each task. Then, integrate the feasible matching sets of all tasks to construct a robot list. ,in For the task The corresponding feasible matching set, the specific feasible matching set That is, from the list of feasible pairs Filter out all items related to the task A collection of paired robots that are capable of performing actions.

[0021] As a further improvement to this technical solution, a mission exclusivity constraint is introduced during the aerospace manufacturing process. Construct a policy set containing the same number of tasks as the task package; iterate through the robot list A, and if the number of tasks in robot list A is... The feasible matching set contains only one robot. Then the task and robots Assign to the policy set; then update robot list A to the robot list. Robot list Each feasible match set in the dataset must include at least two robots;

[0022] Traverse the list of robots Filter the feasible matching set with the fewest robots, and define the corresponding feasible matching set as the robot list. .

[0023] As a further improvement to this technical solution, the branch iterative allocation method is used for the robot list. Each task is assigned to a corresponding robot:

[0024] Select the robot list in sequence For each feasible matching set, if the selected feasible matching set is a feasible matching set... When, if a feasible matching set It includes two robots, namely robot and robots And feasible matching set Corresponding task :

[0025] Forced robot allocation Execute the task The robot employs a dual-reward function adaptation algorithm. Execute the task Time Adaptation Score ;

[0026] The screening includes robots. Multiple feasible matching sets, and the tasks corresponding to each feasible matching set, are assigned to robots one by one. Assign each task and calculate the robot's performance during the assignment process. The adaptation score when performing each corresponding task; filtering for the highest adaptation score. Add up the adaptation score and maximum fit score , to obtain the total score ;

[0027] Similarly, forced allocation of robots Execute the task Calculate the adaptation score and the corresponding maximum fit score , get the total score ;

[0028] compare the total score and the total score : if the total score > the total score , determine to use the robot to perform the task .

[0029] As a further improvement of the technical solution, the double profit function fitting algorithm in step S2 calculates the fitting score of the robot performing the task , first constructs the target profit function corresponding to the robot , and the task profit function corresponding to the task , adds the target profit function and the task profit function to obtain the fitting score . As a further improvement of the technical solution, the target profit function of the robot is constructed as follows:

[0030] ;

[0031] ;

[0032] wherein:

[0033] is the efficiency weight, is the energy consumption weight, and the sum of the efficiency weight and the energy consumption weight is 1; is the standard man-hour of the task ; is the actual man-hour of the robot to complete the task ; is the unit energy consumption coefficient of the robot .

[0034] As a further improvement of the technical solution, the task profit function of the task is constructed as follows:

[0035] ;

[0036] wherein:

[0037] is the task​​​​ Time constraint priority; For robots Complete the task Actual working hours; For the task Standard working hours; For the task The minimum accuracy requirement; For robots The processing accuracy threshold; For the task The accuracy requirements.

[0038] As a further improvement to this technical solution, step S3: obtaining the robot list The system retrieves the robots corresponding to each task in the feasible matching set, and then uses the robot corresponding to each task as the search criteria to filter the robot list for each task. The corresponding task in the selection; if the selected task has been matched by the adaptation score calculated in step S2, then mark the corresponding task; at the same time, count the total number of tasks corresponding to all robots and the number of marked tasks, and calculate the matching ratio of each feasible matching set;

[0039] The matching percentages of all feasible matching sets are retrieved sequentially, and the feasible matching set with the largest matching percentage is selected as the feasible matching set for the branch iterative allocation method to be executed in step S2.

[0040] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0041] In the active control method of the special robot processing system that combines game optimization strategy, the robot can be matched with each task by judging whether the robot has the strategy feasibility to select tasks, forming corresponding matching pairs. Multiple matching pairs are used to construct a feasible pairing list. Then, all robots corresponding to each task are filtered, and the task of the unique robot is assigned to the strategy set first. This avoids robot resource idleness and task execution conflict caused by chaotic task allocation, and ensures that each robot focuses on only a single task.

[0042] After allocation, a robot list C is constructed, and a branch-based iterative allocation method is used to assign corresponding robots to the tasks in the feasible matching set of robot list C:

[0043] The branch iteration allocation method sequentially selects each robot in the feasible matching set, forces the robot to be allocated to the corresponding task, takes the robot forced to be allocated as a search core, screens out all other feasible matching sets containing the robot and the corresponding tasks in the robot list, and then calculates the adaptation score of each robot for the corresponding task by using a double profit function (a target profit function quantifying the efficiency-energy consumption comprehensive performance and a task profit function quantifying the time limit-precision comprehensive performance), so as to realize optimal decision-making in the multi-target coupling scene (such as a robot with high processing efficiency but slightly exceeding the energy consumption, and another robot with energy consumption compliance but slightly longer time limit), and take into account the processing efficiency, energy consumption cost, time limit requirement and precision standard, to meet the stringent requirements of aerospace manufacturing.

[0044] On the basis of the double profit function, the order of selecting the feasible matching set in the robot list C is adjusted again by the tasks and robots with the calculated adaptation scores, the matching proportion of each feasible matching set is calculated, that is, the ratio of the number of tasks in the feasible matching set to the total number of corresponding tasks, so as to quantify the adaptation coverage and concentration of the feasible matching set, and then the feasible matching set with high matching proportion is preferentially selected for processing according to the order from high to low of the matching proportion, which can not only avoid the delay of allocation decision caused by too many feasible matching sets, but also ensure that the robot-task combination with the highest matching value is preferentially used, so as to greatly improve the resource utilization efficiency and task allocation efficiency while ensuring the accuracy of task allocation.

[0045] In addition to the purposes, features and advantages described above, the present application has other purposes, features and advantages. The present application will be further described in detail below with reference to the drawings. BRIEF DESCRIPTION OF DRAWINGS

[0046] Fig. 1 The matching pair working principle diagram is constructed in step S2 of the present application;

[0047] Fig. 2 The working principle flowchart of the branch iteration allocation method in the present application;

[0048] Fig. 3 The overall working principle flowchart of the present application. DETAILED DESCRIPTION

[0049] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0050] The efficient collaborative processing of multiple robots in the additive manufacturing process of aerospace components usually adopts the means of process capability classification. According to different additive process types such as laser selective melting and electron beam wire deposition, special robots with corresponding process execution capabilities are classified and collected to form a robot set ; and the means of combining structure disassembly and process decomposition are adopted. According to the structural characteristics (such as the curved surface structure of the throat of the combustion chamber and the lattice structure of the satellite support) of the aerospace components and the processing requirements, they are disassembled into multiple tasks with clear process requirements and processing range to form a corresponding task package set (such as the throat area of the combustion chamber , the lattice area of the satellite support ), specifically:

[0051] The working principle of dividing the robot set is as follows:

[0052] First, the process capability of all special robots participating in aerospace additive manufacturing is surveyed, and the core parameters of each robot are collected, including but not limited to the following parameters:

[0053] Additive process type (such as laser selective melting, electron beam wire deposition, and electric arc additive manufacturing);

[0054] Processing material adaptability (such as the processing stability of materials such as titanium alloy, high-temperature alloy, and aluminum alloy);

[0055] Processing precision threshold (such as the dimensional tolerance and surface roughness of the formed parts, etc. micron-level indicators);

[0056] Maximum processing range (such as the length, width, and height dimensions of the forming space to adapt to different specifications of aerospace components)

[0057] Then, based on the core parameters, the robots are divided into different subgroups according to the multi-dimensional tag system of process type + material + precision + range, and finally collected to form a robot set , for example:

[0058] All robots with "laser selective melting process, titanium alloy processing capability, ±0.1mm precision, 1m×1m×1m processing range" are classified into subgroup, and so on.

[0059] The working principle of dividing the task package set is as follows:

[0060] First, the aerospace components (such as rocket combustion chambers and satellite supports) are three-dimensionally modeled and structurally analyzed to identify regions with different geometric characteristics (such as curved surfaces, lattices, and thin walls) and mechanical requirements.

[0061] According to the characteristics of the additive manufacturing process, the components are disassembled into multiple task packages according to the principles of process compatibility + structural independence + task size balance, for example:

[0062] The throat area of the rocket combustion chamber requires high sealing due to the complex curved surface, and the laser selective melting process is used to disassemble it into task ; The lattice area of the satellite support needs to be lightweight due to the dense structure, and is disassembled into task ;

[0063] For each disassembled task package, the following standardized task parameters are included but not limited to:

[0064] Process requirements (such as must use electron beam fusion deposition process);

[0065] Material type (such as specified as a certain brand of high-temperature alloy);

[0066] Space requirements (such as the size of the throat area of the combustion chamber , and the space geometry characteristics)

[0067] Precision index (such as size tolerance ≤ ± 0.05mm, surface roughness ≤ Ra1.6μm);

[0068] Time limit requirement (such as need to be completed within 72 hours);

[0069] Finally form a task package set .

[0070] Referring to Figs. 1-3 , the embodiment of the present application provides an active control method of a special robot processing system combined with a game optimization strategy, specifically:

[0071] Step S1, determine whether the robot has the strategy feasibility of selecting the task, so as to match each task corresponding robot, form a corresponding matching pair, and jointly construct a feasible pairing list from multiple matching pairs;

[0072] Step S2, receive the feasible pairing list, filter all robots corresponding to each task in the list, then construct a robot list A, assign the task of the only robot in the robot list A to the strategy set, and update the robot list A to a robot list B;

[0073] Filter the smallest number of feasible matching sets in the robot list B, define the corresponding feasible matching set as a robot list C; use the branch iteration distribution method to assign the corresponding robot to the task of the feasible matching set in the robot list C: forcibly assign the feasible matching set robot in the robot list C, and a double-reward function is used to calculate the adaptation score of each robot corresponding to the task, and the robot with the highest total score is selected to execute the feasible matching set corresponding tasks, and the robots and tasks in other feasible matching sets that have been allocated are removed, and then the robot list C is iteratively updated again until all tasks in the strategy set are allocated;

[0074] Step S3, obtain the robot list C, call out the robots corresponding to the tasks in each feasible matching set, and screen the robot list C with the robots as the search condition corresponding tasks; if the adaptation score of the screened task has been calculated in step S2, the corresponding task is marked; the total number of tasks corresponding to all robots and the number of marked tasks are counted, the matching proportion of each feasible matching set is calculated; the matching proportions of all feasible matching sets are called out in turn, and the feasible matching set with the largest matching proportion is selected as the feasible matching set for the branch iteration allocation method in step S2.

[0075] In the implementation process of the above embodiment, the core parameters such as the process type and material adaptability of the special robot are investigated, the task package is disassembled in combination with the structure of the aerospace component and the processing demand, the robot set and the task package set are formed; then, the feasible pairing list is constructed according to the adaptability of the processing precision, the maximum processing range of the robot and the task precision, the spatial demand; subsequently, the tasks with only one suitable robot are screened and allocated through the task exclusivity constraint, and the branch iteration allocation method is used for the tasks with multiple robot options, and a double-reward function (integration of efficiency, energy consumption, construction period, precision) is used to calculate the adaptation score, and the robot with the optimal total score is selected; finally, the matching proportion of the feasible matching set is calculated to determine the allocation priority, and the combination with the highest adaptation coverage is processed preferentially until all tasks are allocated.

[0076] The feasible pairing list is constructed through the precise matching of the core parameters (processing precision, maximum processing range) of the robot and the task demand (precision, spatial requirement), the tasks with only one suitable robot are allocated preferentially, which can not only avoid the idle of the robot resources and the execution conflict of multiple tasks competing for the same equipment, but also rely on the task exclusivity constraint to eliminate the quality problems such as inconsistent component size and uneven surface quality caused by parameter difference and operation deviation when multiple robots process the same task;

[0077] The method employs a branch-iterative allocation approach to focus on complex scenarios with multiple robot options. It integrates efficiency, energy consumption, schedule, and accuracy metrics using a dual-benefit function to calculate the adaptation score, achieving optimal decision-making in multi-objective coupled scenarios and meeting the stringent requirements of aerospace manufacturing. By calculating the matching ratio of feasible matching sets, the allocation priority is determined, prioritizing combinations with the highest adaptation coverage and concentration. This significantly reduces the computational load of traditional enumeration methods, shortens the task allocation cycle, and improves resource utilization efficiency and allocation accuracy. It effectively solves key problems in additive manufacturing of aerospace components, such as task-robot matching combination explosion, allocation cycle exceeding time, and difficulty in achieving optimal matching for multiple objectives.

[0078] In one embodiment, the specific implementation process of step S1 above is as follows:

[0079] Obtain a set of robots with different additive manufacturing processes after division. And the collection of multiple mission packages formed after disassembling aerospace components. ; respectively call up the robot sets The core parameters and task package set of each robot For each task, the robot and task are paired using core parameters and task parameters to form matching pairs. Then, multiple matching pairs are used to construct a feasible pairing list. Specifically:

[0080] Receive task Precision requirements Space requirements and robots Machining accuracy and maximum processing range ;

[0081] Judge the robot Does it have the ability to select tasks? The feasibility of the strategy is assessed; if the strategy is feasible, then a robot is formed. With the task Corresponding matching pairs and multiple matching pairs Together they form a list of feasible pairs. The specific working principle is as follows:

[0082] If the robot Machining accuracy ≥Task package Precision requirements ,robot Maximum processing range ≥Task package Space requirements When, it means the robot Machining accuracy and maximum processing range Able to meet the task Precision requirements Space requirements Therefore, the robot is judged With task selection The feasibility of the strategy, and for robots Assign tasks Conversely, it determines the robot. It does not have the ability to select tasks. The feasibility of the strategy.

[0083] Task in step S1 above Precision requirements Space requirements Derived from aerospace engineering design specifications, component functional requirements, and manufacturing standards, specifically:

[0084] Precision requirements The performance indicators of aerospace components determine the performance of components. For example, the fitting clearance of rocket fuel valves needs to be controlled within 0.008-0.012mm, and the airfoil profile tolerance of turbine blades needs to reach ±0.05mm. These performance indicators are determined through engineering analysis and simulation verification during the spacecraft design phase to ensure the sealing performance, structural strength, and operational efficiency of components under extreme environments.

[0085] Space requirements The geometric dimensions and installation layout of aerospace components, such as the size characteristics of large components like rocket body sections and satellite load-bearing structures, directly determine the processing range that special robots need to adapt to. The geometric dimensions and installation layout are determined during the mission planning stage through 3D modeling, structural analysis, and other means to form specific mission parameter documents, providing a basis for matching robot capabilities and allocating tasks.

[0086] In summary:

[0087] Step S1 obtains a set of robots with different additive manufacturing processes and a set of task packages formed by disassembling aerospace components. It retrieves the robot's core parameters (additive manufacturing process type, processing accuracy, maximum processing range, material compatibility, etc.) and task parameters (accuracy requirements, space requirements, process requirements, material type, etc.). The feasibility of the strategy is judged by the criteria of robot processing accuracy ≥ task accuracy requirements and maximum processing range ≥ task space requirements. Each task is matched with the corresponding robot to form a matching pair, and finally a feasible pairing list is constructed. This technical means first solves the technical problem of the ambiguity of the basic compatibility between robots and tasks in the additive manufacturing scenario of aerospace components.

[0088] By accurately comparing parameters (such as matching micron-level processing accuracy with task accuracy requirements, and matching component space size with robot processing range), robot-task combinations without basic execution capabilities are eliminated from the source (such as a robot with a processing accuracy of 0.2mm cannot match a task with an accuracy requirement of 0.1mm), thus avoiding invalid combinations occupying computing resources in the subsequent allocation stage.

[0089] Secondly, step S1 solves the problem of robot-task mismatch across processes and materials by pre-screening based on process type and material compatibility (e.g., matching a laser selective melting robot with a task that meets the corresponding process requirements, or a titanium alloy processing robot with a task that matches a titanium alloy component). This prevents the interruption of the processing flow due to the mismatch between the robot's process capabilities and the task requirements (e.g., an arc additive manufacturing robot cannot complete an electron beam filament deposition task). At the same time, the feasible pairing list constructed in step S1 provides a matching sample library with verified basic compatibility for the subsequent step S2, avoiding the need for step S2 to re-screen from all robots and all tasks, greatly reducing the initial data processing volume of step S2, and achieving efficient connection between steps S1 and S2.

[0090] This embodiment further considers that, in scenarios where multiple robots can be adapted to the same task simultaneously, to avoid the problems caused by traditional enumeration methods that require traversing all "task-robot" strategy combinations (the number of combinations grows exponentially with the number of tasks and robots; for example, 12 tasks and 10 robots can form 120 basic combinations, and collaborative scenarios can exceed 1,000), and that each combination needs to independently calculate multi-dimensional indicators such as processing efficiency, energy consumption cost, and accuracy stability, and verify the hard constraints of aerospace manufacturing (verification of a single group requires calling the process database to calibrate indicators and retrieving equipment data across systems to verify constraints, taking tens of seconds per group), the overall computational load far exceeds the system's carrying capacity (industrial computers need to continuously operate for 8-12 hours), the allocation cycle is greatly extended (far exceeding the 24-hour time limit requirement of aerospace missions), and even the optimal execution window of the task is missed due to excessively long enumeration time, therefore:

[0091] In one embodiment, the specific implementation process of step S2 above is as follows:

[0092] Receiving step S1 includes a list of feasible pairings with multiple matching pairs. ;

[0093] For each task, select robots capable of performing the task to form a feasible matching set for each task. Then, integrate the feasible matching sets of all tasks to construct a robot list. ,in For the task The corresponding feasible matching set, the specific feasible matching set That is, from the list of feasible pairs Filter out all items related to the task A collection of paired robots that are capable of performing actions;

[0094] Introducing mission exclusivity constraints in aerospace manufacturing processes Construct a policy set containing the same number of tasks as the task package; use task exclusivity constraints. To ensure that when building a set of policies, a single task Only one robot is assigned to avoid multiple robots handling the same task. At that time, due to the robot Problems such as inconsistent component dimensions and uneven surface quality caused by parameter differences and deviations in operating procedures are addressed, and all tasks are covered without omission.

[0095] Iterate through the robot list A. If the task in robot list A is... The feasible matching set contains only one robot. This indicates the task. Only robots Execution, therefore the task and robots Assign tasks to a set of policies, thereby providing a set of policies for tasks within that set. Assign the only feasible robot And remove other feasible matching sets. ( The robot in ) and tasks Update robot list A to the robot list. Robot list Each feasible match set in the list includes at least two bots, through the bot list. In the subsequent process of assigning a corresponding robot to each task using the branch iterative allocation method, the branch iterative allocation method can be focused on complex matching scenarios with multiple robots available.

[0096] Traverse the list of robots Filter the feasible matching set with the fewest number of robots (the fewest number of robots is greater than 1), and define the corresponding feasible matching set as the robot list. (List of robots) The number of feasible matching sets is (And all feasible matching sets have the same number of robots), and then the branch iterative allocation method is used to assign a corresponding robot to each task. The specific working principle of the branch iterative allocation method is as follows:

[0097] Select the robot list in sequence For each feasible matching set, if the selected feasible matching set is a feasible matching set... When, if a feasible matching set It includes two robots, namely robot and robots And feasible matching set Corresponding task :

[0098] Forced robot allocation Execute the task The robot employs a dual-reward function adaptation algorithm. Execute the task Time Adaptation Score ;

[0099] The screening includes robots. Multiple feasible matching sets, and the tasks corresponding to each feasible matching set, are assigned to robots one by one. Assign each task and calculate the robot's performance during the assignment process. The adaptation score when performing each corresponding task; to ensure the robot The individual contributes the highest single-match value in the task allocation process, and the highest fit score is selected. Add up the adaptation score and maximum fit score , to obtain the total score ;

[0100] Similarly, forced allocation of robots Execute the task Calculate the adaptation score and the corresponding maximum fit score , to obtain the total score ;

[0101] Compare total scores and total score If the total score > Total Score This indicates that the robot Execute the task Therefore, it was determined that a robot would be used. Execute the task ;

[0102] Define the task Corresponding robot Then, tasks were reassigned. and robots To the strategy set, and remove robots from other feasible matching sets. And tasks Then iterates and updates again until all tasks in the strategy set have been assigned.

[0103] By eliminating robots within other feasible matching sets And tasks Ensure the robot list The number of robots in multiple feasible matching sets is gradually reduced, thereby continuously simplifying the problem complexity in subsequent matching processes. For example:

[0104] When the task The feasible matching set for (rocket fuel valve processing) contains only robots. At that time, the task With robots After being assigned to the policy set, other feasible matching sets (such as tasks) are removed. , Robots in the feasible matching set and tasks At this point, the robot list gradually updates from the initial multi-task, multi-robot matching scenario to a list containing only tasks available to multiple robots (such as remaining tasks). , Each feasible matching set contains at least two robots.

[0105] This process gradually locks down the uniquely matching task-robot combination, ensuring the strict satisfaction of the space mission's exclusivity constraint (only one robot executes a single task); and by continuously reducing the number of robot options in subsequent matching scenarios, the subsequent branching iterative allocation method can focus on core scenarios with multiple robot options, such as when the remaining tasks... When the feasible matching set for (satellite bracket lattice processing) is gradually reduced from the initial 5 robots to 2 robots, the computational load of calculating the dual-benefit function (integrating processing efficiency, energy consumption, and accuracy stability) using the branch-iteration allocation method will be significantly reduced. The scenario that originally required enumerating 10 combinations can be simplified to comparing only the total scores of 2 forced allocation schemes. This not only improves matching efficiency (reducing industrial computer processing time from 8-12 hours to several hours), but also enables matching in multi-objective coupled scenarios (such as robots). High processing efficiency but slightly excessive energy consumption, robot (Energy consumption compliance but slightly longer construction period) achieves optimal decision-making through quantitative total score.

[0106] For the list of robots If the number of robots in all feasible matching sets is greater than two, then the corresponding task In the supplementary step S2, the task is allocated using the branch-iteration allocation method. The working principle of the corresponding robot:

[0107] If the robot list In the entire feasible matching set, there are three robots, corresponding to the task. When the time comes, each robot is selected sequentially to be forcibly assigned to perform the corresponding task. ; the remaining robots that have not been selected, screening a plurality of corresponding feasible matching sets, using a double-revenue function fitting algorithm to calculate the fitting score of the robot and each feasible matching set corresponding to the task, and screening the maximum fitting score, and finally comparing to determine the task corresponding to the robot.

[0108] The double-revenue function fitting algorithm in step S2 calculates the fitting score of the robot when performing the task , first constructs the target revenue function of the robot , and the task corresponding to the task revenue function , adds the target revenue function , the task revenue function , and the task revenue function , and the task revenue function to obtain the fitting score , and the specific working principle is:

[0109] The target revenue function of the robot is constructed , which is used to quantify the comprehensive performance of efficiency-energy consumption of the robot when completing the task , and the specific expression is:

[0110] ;

[0111] Among them:

[0112] is the efficiency weight, is the energy consumption weight, and the sum of the efficiency weight and the energy consumption weight is 1;

[0113] is the standard working hours of the task (such as , which can be obtained based on the aerospace process database);

[0114] is the actual working hours of the robot to complete the task , which is determined by the task complexity and the robot capability matching degree, specifically: , wherein is the maximum complexity that the robot can handle, which can be set by statistically analyzing the highest structural complexity (such as the comprehensive quantitative value of indicators such as lattice structure density, surface curvature change, and thin-walled structure ratio) of the aerospace components tasks that the robot has successfully completed in the past, combined with hardware performance redundancy (10-20% performance margin is reserved to cope with processing fluctuations) to set the maximum complexity ; The complexity impact coefficient can be set based on the technological characteristics and error control requirements of aerospace component processing. It is determined by analyzing the correlation between task complexity and actual time deviation in a large amount of task processing data, prioritizing the balance between processing accuracy and efficiency. The value needs to be verified through multiple sets of tasks (e.g., selecting aerospace sub-tasks with different complexities and testing the robot under different complexity impact coefficients). (The degree of consistency between actual working hours and standard working hours). The degree of matching between robot capabilities and task complexity; if the degree of matching... The smaller the value, the higher the efficiency gain for the robot;

[0115] For robots The unit energy consumption coefficient (e.g., 2.5 kJ / h, reflecting the energy consumption control requirements of aerospace manufacturing);

[0116] Used to quantify energy consumption costs and avoid robots In pursuit of efficiency, excessive energy consumption is not in line with the resource management requirements of aerospace manufacturing.

[0117] Build task Task reward function Task reward function Used for quantification tasks Robot The overall performance of schedule and accuracy during execution ensures that the space mission meets accuracy requirements while avoiding delays in key milestones (prioritizing schedules for urgent missions). The specific expression is as follows:

[0118] ;

[0119] in:

[0120] For the task Time constraint priority, time constraint priority The larger the value, the more significant the task. The more urgent;

[0121] For robots Complete the task Actual working hours;

[0122] For the task Standard working hours;

[0123] For the task Minimum accuracy requirements (such as minimum accuracy requirements) );

[0124] the processing precision threshold of the robot ;

[0125] the precision requirement of the task ;

[0126] Ensure that the robot precision is not lower than the task bottom line requirement, the larger the ratio, the better the precision adaptability, avoid the scrap of aerospace components due to insufficient precision, and protect the processing quality of aerospace components.

[0127] Step S2 first uses the branch iteration assignment method to forcibly assign the robot to the corresponding task, and after the assignment is completed, the adaptation score between each robot and different tasks is calculated in turn through the double benefit function adaptation algorithm, and then the optimal matching robot is selected for each task; In the process of forced assignment of branch iteration assignment method, the core logic is to match the robots that have not been assigned to the tasks with strategic feasibility one by one, and the "from few to many" assignment logic is adopted - first process the task with the least number of robots in the feasible matching set, and gradually reduce the range of optional robots for each task;

[0128] Although the branch iteration assignment method can continuously reduce the number of candidate robots corresponding to each task through iteration, there is still an assignment bottleneck: in aerospace engineering, most special robots are pre-calibrated with core capability parameters such as processing precision threshold and maximum processing range. The core capability parameters generally meet the basic precision requirements and space requirements of aerospace tasks (such as the processing precision of multiple robots ≥ the micron-level precision requirement of the task, and the maximum processing range ≥ the space size requirement of the task), and the complexity, time constraint priority and other characteristic parameters of the aerospace task and the capability adaptability of multiple robots are in the effective interval, so there must be a situation where the same task is adapted to multiple robots.

[0129] In summary:

[0130] Step S2 forms a feasible matching set by receiving the feasible pairing list output by step S1, and selects the robots with execution capability for each task to form a feasible matching set and construct a robot list A. The task allocation of the only adapted robot is screened out by introducing the exclusivity constraint of aerospace manufacturing task to the strategy set and updating to robot list B. The feasible matching set with the least number of robots is defined as robot list C by traversing list B. The optimal robot is screened out by using the branch iteration assignment method combined with the double benefit function to calculate the adaptation score and total score, and the assigned robots and tasks are removed to update list C until all tasks are assigned.

[0131] In the traditional scenario, 12 tasks and 10 robots form 120 combinations, and in the collaborative scenario, there are more than 1,000 combinations. The multi-dimensional indicators of each combination need to be verified independently, which takes 8-12 hours. In step S2, the robots with only execution capability are filtered out from the list A to eliminate invalid combinations. Then, the unique matching task is locked from the list B to reduce the subsequent decision-making amount. Finally, the focus is on the smallest feasible matching set in the list C (e.g., 2 robots for 1 task), so that the branch iteration distribution method only needs to compare the total scores of 2 groups (instead of all combinations). At the same time, the allocated resources are continuously eliminated during iteration, reducing the calculation amount from exponential to linear, greatly shortening the operation time of industrial computers to a few hours, meeting the 24-hour distribution period requirement of space tasks, and avoiding missing the best execution window.

[0132] Secondly, for the multi-objective coupling decision problem of multi-robot selectable tasks, step S2 converts the dispersed efficiency (actual working hours), energy consumption (unit energy consumption coefficient), time limit (time constraint priority), and precision (processing precision threshold) dimensions into quantifiable matching scores through double profit functions (target profit function quantifies the efficiency-energy consumption comprehensive performance, and task profit function quantifies the time limit-precision comprehensive performance). Then, combined with the forced distribution of the branch iteration distribution method and the comparison of the total scores (e.g., the score of the robot executing the current task and the sum of the maximum scores of other tasks), the coupling constraints between dimensions (e.g., the contradiction between improving efficiency and controlling energy consumption) are balanced, and the global optimal decision is achieved, avoiding the subjectivity of traditional experience-based decision-making.

[0133] The task of the unique matching robot is forced to be distributed through the task exclusivity constraint, and it is clear that only one robot executes a single task, which fundamentally eliminates the quality risks such as inconsistent component size and uneven surface roughness caused by parameter differences (e.g., processing precision threshold deviation) and operation deviation when multiple robots process the same task, ensuring the micron-level tolerance requirement of space components. In addition, step S2 clearly divides the task distribution priority (first unique matching task, then multiple robots with fewer tasks) through layer-by-layer screening and updating from list A to list C, avoiding resource idling (e.g., unique matching robots not being allocated in time) and execution conflicts (multiple tasks competing for the same robot), and improving distribution efficiency. Meanwhile, step S2 takes the feasible matching list output by S1 as the starting point, without the need to re-verify the basic adaptability of robots and tasks (e.g., whether the processing precision and space requirements meet the standards).

[0134] Further, the robot list If not screened specifically, multiple feasible matching sets in the robot list will result in low efficiency of task-robot distribution. Therefore, to avoid the above problems:

[0135] In one embodiment, the specific implementation process of the above step S3 is as follows:

[0136] Obtaining a robot list , and calling out the robots corresponding to each task in each feasible matching set, and then taking the robots corresponding to each task as the retrieval condition to screen the tasks corresponding to the robots in the robot list ; if the screened tasks have been calculated for the adaptation score in step S2, the corresponding tasks are marked; meanwhile, the total number of tasks corresponding to all robots and the number of marked tasks are counted, and the matching proportion of each feasible matching set (the number of marked tasks / the total number of corresponding tasks) is calculated; the matching proportions of all feasible matching sets are called out in turn, and the feasible matching set with the largest matching proportion is selected as the feasible matching set for the branch iteration assignment method in step S2.

[0137] Therefore, by quantifying the matching proportion of the feasible matching set, in the scenario where multiple feasible matching sets coexist, the feasible matching set with the highest coverage and concentration of task-robot adaptation score is quickly identified, providing clear priority for the assignment in step S2; not only is the delay of the assignment decision caused by too many feasible matching sets avoided, but also the task-robot combination with the highest matching value is ensured to be assigned preferentially, thereby improving the overall assignment efficiency and the accuracy of resource utilization in the multi-robot adaptation scenario of space tasks.

[0138] In summary: steps S1, S2 and S3 are progressively matched in the form of basic adaptation screening-multi-objective optimization assignment-priority dynamic calibration, forming a full-process closed-loop control system to efficiently solve the core problem of multi-robot collaborative processing in space component additive manufacturing: step S1 constructs a feasible pairing list of basic adaptation qualification by quantitatively matching the core parameters of robots and task requirements, thereby reducing the initial data processing amount for steps S2 and S3 and avoiding the consumption of invalid computing power; step S2 takes the list to lock the only adaptation task with task exclusivity constraint, focuses on the multi-robot scenario with the branch iteration assignment method, quantifies the multi-objective score with the double-income function, and updates the list after excluding the assigned resources to provide dynamic data and adaptation score support for step S3, thereby avoiding assignment faults; step S3 calculates the matching proportion based on the results of step S2, reversely optimizes the assignment order of step S2, preferentially processes combinations with high adaptation coverage, and helps step S2 to avoid combination explosion and global resource waste.

[0139] The basic principles, main features and advantages of the present application are shown and described above. Those skilled in the art should understand that the present application is not limited by the above examples, and the above examples and descriptions in the specification are only preferred examples of the present application and are not intended to limit the present application. Without departing from the spirit and scope of the present application, various changes and improvements can be made to the present application, and these changes and improvements all fall within the scope of the claimed present application. The scope of protection of the present application is defined by the appended claims and their equivalents.

Claims

1. An active control method for a special robot processing system combining game-theoretic optimization strategies, characterized in that, The work steps are as follows: S1. Determine whether the robot has the strategy feasibility to select tasks, thereby matching the robot corresponding to each task to form corresponding matching pairs, and constructing a feasible pairing list by combining multiple matching pairs. S2. Receive the list of feasible pairings, filter all robots corresponding to each task, then construct robot list A, assign the task of constructing the unique robot in robot list A to the policy set, and update robot list A to robot list B. Filter the feasible matching set with the fewest robots from robot list B, and define the corresponding feasible matching set as robot list C; use the branch-iterative allocation method to assign corresponding robots to the tasks in the feasible matching set of robot list C: force allocation of feasible matching set. The system identifies robots and uses a dual-reward function to calculate the adaptation score for each robot's corresponding task. Then, it calculates the total score based on the adaptation scores and selects the robot with the highest total score to perform feasible matching. The corresponding task is identified; and other robots and tasks already assigned in the feasible matching set are removed. Then, the robot list C is updated iteratively again until all tasks in the strategy set have been assigned. S3. Obtain robot list C, retrieve the robot corresponding to each task in the feasible matching set, and filter the robot list by robot as the search condition. The corresponding task in the selection; if the selected task has been matched by the adaptation score calculated in step S2, then mark the corresponding task; count the total number of tasks corresponding to all robots and the number of marked tasks, and calculate the matching percentage of each feasible matching set; The matching percentages of all feasible matching sets are retrieved sequentially, and the feasible matching set with the largest matching percentage is selected as the feasible matching set for the branch iterative allocation method to be executed in step S2.

2. The active control method for a special robot processing system combining game-theoretic optimization strategies according to claim 1, characterized in that: Step S1 obtains a set of robots with different additive manufacturing processes after division. And the collection of multiple mission packages formed after disassembling aerospace components. ; respectively call up the robot sets The core parameters and task package set of each robot For each task, the robot and task are paired using core parameters and task parameters to form matching pairs. Then, multiple matching pairs are used to construct a feasible pairing list. .

3. The active control method for a special robot processing system combining game-theoretic optimization strategies according to claim 1, characterized in that: Step S1 constructs matching pairs Receive task Precision requirements Space requirements and robots Machining accuracy and maximum processing range ; Judge the robot Does it have the ability to select tasks? Feasibility of the strategy: If the robot Machining accuracy ≥Task package Precision requirements ,robot Maximum processing range ≥Task package Space requirements When, then determine the robot With task selection The feasibility of the strategy, and for robots Assign tasks Conversely, it determines the robot. It does not have the ability to select tasks. The feasibility of the strategy.

4. The active control method for a special robot processing system combining game-theoretic optimization strategies according to claim 3, characterized in that: Step S2 receives a list of feasible pairs that contains multiple matching pairs from step S1. ; For each task, select robots capable of performing the task to form a feasible matching set for each task. Then, integrate the feasible matching sets of all tasks to construct a robot list. ,in For the task The corresponding feasible matching set, the specific feasible matching set That is, from the list of feasible pairs Filter out all items related to the task A collection of paired robots that are capable of performing actions.

5. The active control method for a special robot processing system combining game-theoretic optimization strategies according to claim 4, characterized in that: Introducing mission exclusivity constraints in aerospace manufacturing processes Construct a policy set containing the same number of tasks as the task package; iterate through the robot list A, and if the task in robot list A is... The feasible matching set contains only one robot. Then the task and robots Assign to the policy set; then update robot list A to the robot list. Robot list Each feasible match set in the dataset must include at least two robots; Traverse the list of robots Filter the feasible matching set with the fewest robots, and define the corresponding feasible matching set as the robot list. .

6. The active control method for a special robot processing system combining game-theoretic optimization strategies according to claim 5, characterized in that: The branch-iterative allocation method is used to generate the robot list. Each task is assigned to a corresponding robot: Select the robot list in sequence For each feasible matching set, if the selected feasible matching set is a feasible matching set... When, if a feasible matching set It includes two robots, namely robot and robots And feasible matching set Corresponding task : Forced robot allocation Execute the task The robot employs a dual-reward function adaptation algorithm. Execute the task Time Adaptation Score ; The screening includes robots. Multiple feasible matching sets, and the tasks corresponding to each feasible matching set, are assigned to robots one by one. Assign each task and calculate the robot's performance during the assignment process. The adaptation score when performing each corresponding task; Filter by maximum fit score Add up the adaptation score and maximum fit score , to obtain the total score ; Similarly, forced allocation of robots Execute the task Calculate the adaptation score and the corresponding maximum fit score , to obtain the total score ; Compare total scores and total score If the total score > Total Score Then it is determined that a robot is used. Execute the task .

7. The active control method for a special robot processing system combining game-theoretic optimization strategies according to claim 6, characterized in that: The dual-reward function adaptation algorithm described in step S2 is used to compute the robot. Execute the task Time Adaptation Score First, build the robot Corresponding target return function and tasks Corresponding task reward function Adding the objective return function Task reward function Get the fit score .

8. The active control method for a special robot processing system combining game-theoretic optimization strategies according to claim 7, characterized in that, Building robots The target return function : ； in: For efficiency weighting, Energy consumption weight and efficiency weight and energy consumption weight The sum is 1; For the task Standard working hours; For robots Complete the task Actual working hours; For robots The unit energy consumption coefficient.

9. The active control method for a special robot processing system combining game-theoretic optimization strategies according to claim 8, characterized in that, Build task Task reward function : ； in: For the task Time constraints priority; For robots Complete the task Actual working hours; For the task Standard working hours; For the task The minimum accuracy requirement; For robots The processing accuracy threshold; For the task The accuracy requirements.

10. The active control method for a special robot processing system combining game-theoretic optimization strategies according to claim 9, characterized in that, Step S3: Obtain the robot list The system retrieves the robots corresponding to each task in the feasible matching set, and then uses the robot corresponding to each task as the search criteria to filter the robot list for each task. The corresponding task in the selection; if the selected task has been matched by the adaptation score calculated in step S2, then mark the corresponding task; at the same time, count the total number of tasks corresponding to all robots and the number of marked tasks, and calculate the matching ratio of each feasible matching set; The matching percentages of all feasible matching sets are retrieved sequentially, and the feasible matching set with the largest matching percentage is selected as the feasible matching set for the branch iterative allocation method to be executed in step S2.

Images