Machine vision processing assistance method and system for multi-axis robots

Abstract

This invention discloses a machine vision-assisted processing method and system for multi-axis robots. The method includes determining the priority of tasks to be processed to divide them into multiple processing task units; obtaining configuration parameters of the processing system, including the arrival rate of task loading, the number of robotic arms in the multi-axis robot, and the service rate of each robotic arm; determining a task allocation model based on the configuration parameters when executing any processing task unit; monitoring the robotic arm force and motion trajectory of the processing task unit, and adaptively adjusting the robotic arm force and motion trajectory according to a preset algorithm; and evaluating the processing quality of the current processing task unit in real time after its completion, providing feedback to adjust the task allocation model, and using the updated task allocation model for the next processing task unit. This invention enables multi-axis robot robotic arms to have a more balanced load, significantly improve resource utilization, and improve processing quality while reducing costs.

Description

Technical Field

[0001] This invention relates to the field of robot control technology, and in particular to a machine vision processing assistance method and system for multi-axis robots. Background Technology

[0002] In the field of industrial automation, multi-axis robot machine vision processing systems have been widely used in precision manufacturing, electronic assembly, and automotive parts processing. Traditional multi-axis robot processing systems typically employ fixed task allocation strategies and preset motion trajectory control methods. The working parameters of the robotic arms are set based on human experience, and processing tasks are executed in a predetermined sequence. With the development of intelligent manufacturing technology, some advanced systems have introduced task priority scheduling and simple parallel processing mechanisms, but the following drawbacks still exist: 1) Existing systems mostly use static priority or fixed-sequence task allocation, making it difficult to dynamically respond to changes in order urgency, leading to delays in urgent order processing, affecting overall production efficiency, and resulting in poor flexibility. 2) The lack of a dynamic task allocation process based on real-time system configuration parameters easily leads to some robotic arms operating under overload while others remain idle, resulting in low resource utilization. 3) Traditional systems cannot adaptively adjust operating parameters based on real-time feedback from the processing process, leading to unstable processing accuracy. 4) Quality feedback is delayed; the evaluation of processing quality is usually conducted after the entire batch of tasks is completed, making it difficult to promptly detect processing anomalies and adjust subsequent tasks, increasing the defect rate and production costs. Summary of the Invention

[0003] To address at least one of the aforementioned technical problems, the present invention provides a machine vision-assisted processing method and system for multi-axis robots.

[0004] In a first aspect, the present invention provides a machine vision-assisted processing method for a multi-axis robot, the method comprising:

[0005] The order urgency is determined by the order placement time of the task to be processed. The priority of the task to be processed is determined based on the order urgency. The task to be processed is divided into multiple processing task units according to the priority order. Each processing task unit contains multiple subtasks that can be processed in parallel.

[0006] Obtain the configuration parameters of the processing system, including the arrival rate of the task to be processed, the number of robotic arms of the multi-axis robot, and the service rate of each robotic arm.

[0007] When executing any processing task unit, the task allocation model is determined according to the configuration parameters of the processing system; the force and motion trajectory of the robotic arm of the processing task unit are monitored when the processing system is working, and the force and motion trajectory of the robotic arm are adaptively adjusted according to the preset algorithm;

[0008] Once a processing task unit is completed, the processing quality of the current processing task unit is evaluated in real time. The task allocation model is adjusted based on the processing quality feedback, and the updated task allocation model is used for the next processing task unit.

[0009] Preferably, determining the task allocation model based on the configuration parameters of the processing system when executing any processing task unit includes:

[0010] The urgency level is calculated based on the order placement time, delivery cycle, and task complexity coefficient.

[0011] For the highest priority of urgency, a greedy algorithm is used to directly assign robotic arms, selecting the robotic arm with the highest service rate and an idle rate greater than 50%.

[0012] For tasks with the next highest urgency priority, an improved genetic algorithm is used to assign robotic arms.

[0013] Preferably, the method of assigning robotic arms using an improved genetic algorithm includes:

[0014] Chromosome encoding is performed, and the matching method between the task and the robotic arm is represented in binary.

[0015] Determine the fitness function by calculating the weighted sum of total time and total energy consumption;

[0016] Determine the mutation rules and retain the three best solutions without mutation.

[0017] Preferably, the method further includes:

[0018] The task allocation results are displayed using a two-dimensional matrix. When the actual speed of the robotic arm decreases by 10%, 30% of the tasks of the corresponding robotic arm are automatically transferred to the adjacent robotic arm.

[0019] Preferably, the service rate of each robotic arm is calculated in the following way:

[0020] The basic service rate is calculated based on the total welding time of a robotic arm, tool switching time, and positioning accuracy compensation coefficient.

[0021] The base service rate is adjusted in real time based on the load attenuation coefficient, collaborative waiting loss, and fault degradation factor to obtain a dynamically updated real-time service rate.

[0022] Preferably, when the monitoring and processing system is working, the force and trajectory of the robotic arm of the processing task unit are adaptively adjusted according to a preset algorithm.

[0023] ;

[0024] ;

[0025] In the formula, It is the real-time adjustable force of the robotic arm. It is the initial force of the robotic arm. It is the error between the target force and the actual force. , , These are the proportional coefficient, integral coefficient, and differential coefficient; It is the real-time adjustable movement speed of the robotic arm. It is a Jacobian matrix, representing the relationship between the joint angles of the robotic arm and the position of the end effector. It is the error between the target position and the actual position.

[0026] Preferably, the real-time evaluation of the processing quality of the current processing task unit includes:

[0027] The evaluation indicators are obtained, including the dimensional accuracy, surface quality, defect density and process stability of the processed parts, and the weights of the corresponding indicators are calculated. The processing quality score is calculated by weighted summation of the various evaluation indicators.

[0028] Secondly, the present invention also provides a machine vision processing assistance system for a multi-axis robot, the system comprising:

[0029] The task division unit is used to obtain the order placement time of the task to be processed to determine the order urgency, determine the priority of the task to be processed based on the order urgency, and divide the task to be processed into multiple processing task units according to the priority order. Each processing task unit contains multiple subtasks that can be processed in parallel.

[0030] The parameter acquisition unit is used to acquire the configuration parameters of the processing system, including the arrival rate of the task to be processed, the number of robotic arms of the multi-axis robot, and the service rate of each robotic arm.

[0031] The task allocation unit is used to determine the task allocation model based on the configuration parameters of the machining system when executing any machining task unit; it monitors the force and motion trajectory of the robotic arm of the machining task unit when the machining system is working, and adaptively adjusts the force and motion trajectory of the robotic arm according to a preset algorithm.

[0032] The iterative optimization unit is used to evaluate the processing quality of the current processing task unit in real time after it is completed, adjust the task allocation model based on the processing quality feedback, and apply the updated task allocation model to the next processing task unit.

[0033] Thirdly, the present invention also provides an electronic device including a processor and a memory, the memory being used to store computer program code, the computer program code including computer instructions, wherein when the processor executes the computer instructions, the electronic device performs the method as described in the first aspect above and any possible implementation thereof.

[0034] Fourthly, the present invention also provides a computer-readable storage medium storing a computer program, the computer program including program instructions that, when executed by a processor of an electronic device, cause the processor to perform a method as described in the first aspect above and any possible implementation thereof.

[0035] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0036] 1) This invention calculates urgency by obtaining the order placement time, dynamically determines task priority, and divides processing task units, solving the problem that traditional systems cannot respond to changes in order urgency in real time. Dynamic priority scheduling allows urgent orders to be processed first, shortening the average order processing time; the parallel subtask processing mechanism fully utilizes the collaborative capabilities of multiple robotic arms, improving the task throughput per unit time.

[0037] 2) A task allocation model was established based on the processing system configuration parameters, such as the material arrival rate, the number of robotic arms, and the service rate. This enabled dynamic optimization of robotic arm resources, avoiding resource idleness and overload. The task allocation model based on system configuration parameters resulted in a more balanced load on the robotic arms and a significant improvement in resource utilization.

[0038] 3) By monitoring the force and trajectory of the robotic arm and making real-time adjustments using a preset algorithm, the problem of insufficient adaptability to parameter changes during processing in traditional systems is solved. The real-time monitoring and adaptive adjustment mechanism can compensate for processing errors caused by factors such as changes in material properties and mechanical wear, effectively improving processing accuracy.

[0039] 4) Immediately after the completion of each task unit, the processing quality is evaluated, and feedback is used to adjust the task allocation model, forming a closed-loop control that solves the problem of lagging quality feedback in traditional systems. This quality feedback adjustment mechanism enables the system to correct processing deviations in a timely manner, reducing the defect rate and significantly saving production costs. It also allows the multi-axis robot processing system to quickly adapt to changes in order demand and fluctuations in the production environment, providing technical support for achieving flexible manufacturing.

[0040] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit this disclosure. Attached Figure Description

[0041] To more clearly illustrate the technical solutions in the embodiments of the present invention or the background art, the accompanying drawings used in the embodiments of the present invention or the background art will be described below.

[0042] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this disclosure and, together with the specification, serve to illustrate the technical solutions of this disclosure.

[0043] Figure 1 A flowchart illustrating a machine vision-assisted processing method for a multi-axis robot provided in an embodiment of the present invention;

[0044] Figure 2 for Figure 1 A flowchart illustrating the sub-steps of step S30;

[0045] Figure 3 for Figure 2 A flowchart illustrating the sub-steps of step S303;

[0046] Figure 4 This is a schematic diagram of a machine vision processing assistance system for a multi-axis robot, provided as an embodiment of the present invention. Detailed Implementation

[0047] To enable those skilled in the art to better understand the present invention, 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. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0048] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of the invention. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0049] Please see Figure 1 , Figure 1 This is a flowchart illustrating a machine vision-assisted processing method for a multi-axis robot, provided as an embodiment of the present invention. Figure 1 As shown, the method includes:

[0050] S10. Obtain the order placement time of the task to be processed to determine the urgency of the order. Determine the priority of the task to be processed based on the urgency of the order. Divide the task to be processed into multiple processing task units according to the priority order. Each processing task unit contains multiple subtasks that can be processed in parallel.

[0051] Urgency is calculated based on order placement time, delivery cycle, and task complexity coefficient. All tasks are prioritized based on urgency, and a dynamic programming algorithm is used to generate an optimal priority sequence, ensuring that high-urgency tasks are processed first. Task dependency modeling using graph theory decomposes tasks into multiple independent subtasks, and a greedy algorithm aggregates these subtasks into processing task units, maximizing parallelism within each unit. Through dynamic priority scheduling, urgent order processing time is shortened, and task unitization improves parallel processing efficiency, significantly improving order response speed and equipment utilization.

[0052] S20. Obtain the configuration parameters of the processing system, including the arrival rate of the task to be processed, the number of robotic arms of the multi-axis robot, and the service rate of each robotic arm.

[0053] Photoelectric sensors and RFID tags are deployed on the material conveyor line, and the material arrival rate is calculated in real time using a Poisson process model. A multi-robotic arm service system model is established, with the service rate of each robotic arm calculated from historical processing data. A queuing theory model is used to describe the system characteristics. Preferably, system data can be collected every 5 minutes, and a Kalman filter is used to predict parameter change trends to ensure model accuracy. Accurate system parameter modeling reduces task allocation errors and queuing time.

[0054] S30. When executing any processing task unit, determine the task allocation model according to the configuration parameters of the processing system; monitor the force and motion trajectory of the robotic arm of the processing task unit when the processing system is working, and adaptively adjust the force and motion trajectory of the robotic arm according to the preset algorithm.

[0055] A multi-objective optimization model can typically be used, employing a genetic algorithm to find the optimal allocation scheme, recalculating every 10 seconds. Impedance control theory can be used, with real-time feedback from force sensors and a PID controller to adjust the robotic arm's force. Machine vision is used to detect workpiece position deviations, and curve fitting generates a smooth correction trajectory. The dynamic allocation model improves robotic arm utilization, while the adaptive adjustment mechanism enhances machining accuracy.

[0056] S40. When a processing task unit is completed, the processing quality of the current processing task unit is evaluated in real time. The task allocation model is adjusted according to the processing quality feedback, and the updated task allocation model is used for the next processing task unit.

[0057] After a processing task unit is completed, the processing quality of that unit is evaluated in real time. The task allocation model is then adjusted based on the quality feedback. For example, if the processing quality is good and the score is high, it indicates that the current processing parameters are good and minimal changes are needed. If the processing quality is poor, the tasks are reassigned to accommodate the next processing task unit. Specifically, when adjusting the task allocation model parameters based on the quality evaluation results, if the defect rate exceeds 5%, the priority weight for accuracy is increased; if the processing time is excessively long, the priority weight for efficiency is increased. This closed-loop quality control reduces the defect rate, achieving a balance between efficiency and quality.

[0058] See Figure 2 In one embodiment, determining the task allocation model based on the configuration parameters of the processing system when executing any processing task unit includes:

[0059] S301. Calculate the urgency level based on the order placement time, delivery cycle, and task complexity coefficient;

[0060] Urgency = (Current time - Order time) / Standard delivery cycle × Task complexity coefficient, where the task complexity coefficient can be determined by several indicators such as time constraints, set complexity constraints, resource dependencies, and process variability constraints.

[0061] S302. For the robot arm with the highest urgency priority, a greedy algorithm is used to directly allocate the robot arm, selecting the one with the highest service rate and an idle rate greater than 50%; where idle rate = 1 - current load / maximum load;

[0062] S303. For those with the second highest urgency priority, an improved genetic algorithm is used to assign robotic arms.

[0063] See Figure 3 In one embodiment, the method of assigning the robotic arm using an improved genetic algorithm includes:

[0064] S3031. Perform chromosome encoding to represent the matching method between the task and the robotic arm in binary;

[0065] Line: Represents the task index;

[0066] Column: Represents the index of the robotic arm;

[0067] Values: 1 indicates allocation, 0 indicates no allocation;

[0068] Constraint handling:

[0069] Each line contains only one 1 (the task is assigned to only one robotic arm).

[0070] The number of 1s in each column is less than or equal to the robotic arm's processing capacity;

[0071] S3032. Determine the fitness function, calculated based on the weighted sum of total time and total energy consumption;

[0072] The total processing time is the processing time of each robotic arm, and the total energy consumption is the power of each robotic arm multiplied by the processing time. To make the calculation more accurate, a load imbalance degree, i.e., the standard deviation (load rate of each robotic arm), can be introduced. This is calculated by weighting the total processing time, total energy consumption, and load imbalance degree. Therefore, multi-objective optimization balances efficiency and energy consumption, with dynamic weights adapting to different production scenarios, and the load balance degree index reducing equipment wear and tear.

[0073] S3033. Determine the mutation rules and retain the three best solutions without mutation.

[0074] The three solutions with the highest fitness in each generation are retained and directly enter the next generation. Priority is given to mutating robotic arm assignments with excessive load.

[0075] Increase the mutation probability for tasks nearing their delivery deadline. Adaptive mutation rate balances global search and local optimization, and targeted mutation improves the stability of handling urgent tasks.

[0076] This embodiment uses a greedy algorithm to handle urgent tasks and a genetic algorithm to optimize routine tasks, thus achieving a balance between emergency response and global optimization. It avoids using complex algorithms for all tasks, greatly reduces the overall scheduling computation time, and improves processing efficiency.

[0077] Preferably, the method further includes:

[0078] The task allocation results are displayed using a two-dimensional matrix. When the actual speed of the robotic arm decreases by 10%, 30% of the tasks of the corresponding robotic arm are automatically transferred to the adjacent robotic arm.

[0079] A two-dimensional matrix intuitively displays the task-robotic arm mapping relationship, reducing human decision-making time; a prediction mechanism based on real-time speed monitoring avoids the risk of robotic arm downtime due to overload; and an automatic transfer strategy reduces the impact of single-arm failure on overall production. Simulation verification shows that a 30% task transfer volume represents the optimal balance between efficiency and stability. This extends the equipment's lifespan.

[0080] In a preferred embodiment, determining the task allocation model based on the configuration parameters of the processing system includes:

[0081] Construct task allocation objective functions with the goals of load balancing and minimizing total system latency, respectively:

[0082] ;

[0083] In the formula, Indicates the first The number of tasks for each robotic arm, totaling One robotic arm, This represents the real-time service rate of each robotic arm. The arrival rate of packaged and loaded parts for processing; This indicates the total system wait time;

[0084] Establish constraints:

[0085] ;

[0086] ;

[0087] ;

[0088] ;

[0089] ;

[0090] In the formula, It is system utilization. It is the first The queue length of each robotic arm It is the total number of processed parts in the system; It is the first The waiting time of each robotic arm This is the total system waiting time;

[0091] A genetic algorithm is used to solve the task allocation objective function and determine the task load for each robotic arm.

[0092] In this embodiment, an objective function optimization approach was used to determine the initial optimal task allocation scheme. First, two objective functions were established. The first objective function aims to minimize the difference between the actual processing time and the ideal processing time of each robotic arm, ensuring load balancing. The second objective function aims to minimize the total waiting time in the system, improving overall production efficiency. When constructing constraints, it is necessary to ensure that the system will not be overloaded, and the system utilization rate must be less than 1. It is also necessary to ensure that the queue length of each robotic arm and the total number of processed parts in the system are non-negative. Finally, it is necessary to ensure that the waiting time of each robotic arm and the total waiting time of the system are within a reasonable range.

[0093] Furthermore, the model is solved using a genetic algorithm, including:

[0094] 1) Initialize the population: Randomly generate a set of task allocation schemes, each scheme representing the number of tasks for each robotic arm.

[0095] 2) Fitness Calculation: For each task allocation scheme, calculate the values ​​of two objective functions: the load balancing objective function and the objective function for minimizing the total system latency. The fitness value can be a weighted sum of the two objective function values, or a multi-objective optimization method can be used.

[0096] 3) Selection: Select a subset of individuals to enter the next generation based on their fitness values. Common selection methods include roulette wheel selection and tournament selection.

[0097] 4) Crossover: Perform a crossover operation on the selected individuals to generate new individuals. Common crossover methods include multi-point crossover and uniform crossover.

[0098] 5) Mutation: Mutation operations are performed on newly generated individuals to introduce diversity. Common mutation methods include bit flip mutation and insertion mutation.

[0099] 6) Termination condition: The predetermined number of iterations or fitness value is reached and there is no longer significant improvement.

[0100] Therefore, by minimizing the difference between the actual processing time and the ideal processing time of each robotic arm, the load balance of each robotic arm is ensured, avoiding situations where some robotic arms are overloaded while others are idle, thus improving overall production efficiency. By minimizing the total system latency, the waiting time of processed parts in the system is reduced, improving production efficiency and shortening the production cycle. By establishing reasonable constraints, the system is ensured to operate within a reasonable range, avoiding overload and unnecessary waiting, thereby improving the stability and reliability of the system.

[0101] In one embodiment, the service rate of each robotic arm is calculated in the following manner:

[0102] The basic service rate is calculated based on the total welding time of a robotic arm, tool switching time, and positioning accuracy compensation coefficient.

[0103] The base service rate is adjusted in real time based on the load attenuation coefficient, collaborative waiting loss, and fault degradation factor to obtain a dynamically updated real-time service rate.

[0104] Specifically as follows:

[0105] ;

[0106] ;

[0107] In the formula, The base service rate of a certain robotic arm. For dynamically adjusted real-time service rates; For the standard welding cycle of a certain robotic arm, This is the positioning accuracy compensation coefficient. For tools The switching time is in total. A tool; The total time spent welding a certain robotic arm; These are the load attenuation coefficient, cooperative waiting loss, and fault degradation factor, respectively; satisfying:

[0108] ;

[0109] ;

[0110] ;

[0111] In the formula, For continuous working hours, Number of times to wait for collaboration This represents the percentage of downtime due to malfunctions.

[0112] This service rate calculation model is based on the "layered attenuation" theory, decomposing the robotic arm's service capacity into two parts: basic capacity and dynamic loss. The basic service rate reflects the robotic arm's processing capacity under ideal conditions, and is jointly determined by welding efficiency, tool changeover cost, and positioning accuracy. In the formula, The impact of tool switching on actual work efficiency was quantified. The positioning accuracy compensation coefficient reflects the speed constraint imposed by high-precision tasks and meets the requirements. Service rates declined. The load attenuation coefficient is based on thermodynamic principles. The heat generated by the continuous operation of the robotic arm will cause the motor efficiency to decrease, exhibiting an exponential attenuation characteristic. The concept of collaborative waiting loss is represented by the M / M / c model in queuing theory, where the waiting time during multi-arm collaboration follows the loss-to-waiting-time ratio. The early failure characteristics of the bathtub curve are used to identify the failure degradation factor, and the proportion of downtime due to failure directly reflects the health status of the equipment.

[0113] By using multi-dimensional loss modeling, the service rate prediction error is significantly reduced, the load decay model improves the accuracy of preventive maintenance plans, the collaborative loss factor improves the efficiency of multi-robotic arm collaboration, and the dynamic service rate enables the task allocation system to predict the performance changes of robotic arms 2-3 cycles in advance. This achieves a paradigm shift from "static parameters" to "dynamic performance", which is particularly suitable for intelligent manufacturing scenarios with extremely high requirements for accuracy and reliability.

[0114] In one embodiment, when the monitoring and processing system is working, the force and trajectory of the robotic arm of the processing task unit are adaptively adjusted according to a preset algorithm.

[0115] ;

[0116] ;

[0117] In the formula, It is the real-time adjustable force of the robotic arm. It is the initial force of the robotic arm. It is the error between the target force and the actual force. , , These are the proportional coefficient, integral coefficient, and differential coefficient; It is the real-time adjustable movement speed of the robotic arm. It is a Jacobian matrix, representing the relationship between the joint angles of the robotic arm and the position of the end effector. It is the error between the target position and the actual position.

[0118] In this embodiment, an initial robotic arm force is preset based on the material and diameter of the workpiece. Force sensors installed on the end effector of the robotic arm can monitor the actual force during the machining process in real time. Calculate the error between the target force and the actual force. A PID controller is used to determine the error between the actual force and the target force. Dynamically adjust the force of the robotic arm This can be adjusted based on actual production data and experimental results. , , The size of the target position is used to ensure precise control of the robotic arm's force. Similarly, a visual servo control algorithm is used to adjust the target position based on the error between the target position and the actual position. Dynamically adjust the movement speed of the robotic arm Based on actual production data and experimental results, the parameters of the Jacobian matrix and control algorithm are adjusted to ensure the accuracy and stability of the robotic arm's movement.

[0119] Therefore, this embodiment uses PID control to adjust the robotic arm's force in real time, ensuring that the workpiece is not damaged during processing and improving processing quality. By dynamically adjusting the robotic arm's force, it adapts to workpieces of different materials and sizes, improving the system's flexibility and adaptability. Visual servo control is used to adjust the robotic arm's movement speed in real time, ensuring the accuracy of the processing position and improving packaging accuracy. Dynamically adjusting the robotic arm's movement trajectory avoids excessive movement and collisions, improving system stability and safety, reducing rework rates, shortening the production cycle, and lowering production costs.

[0120] In one embodiment, the real-time evaluation of the processing quality of the current processing task unit includes:

[0121] The evaluation indicators are obtained, including the dimensional accuracy, surface quality, defect density and process stability of the processed parts, and the weights of the corresponding indicators are calculated. The processing quality score is calculated by weighted summation of the various evaluation indicators.

[0122] In this embodiment, the process of obtaining each indicator is as follows:

[0123] Dimensional accuracy: Key dimensional data are acquired through a laser displacement sensor with a sampling frequency of 100Hz and a measurement accuracy of ±0.01mm;

[0124] Surface quality: Surface images are captured by an industrial camera with a resolution of 12 megapixels, and surface roughness is analyzed using a surface roughness algorithm;

[0125] Defect density: The deep learning model detects surface scratches, pores, and other defects with a false detection rate of less than 0.5%;

[0126] Process stability: Collect force / torque data of the robotic arm and analyze the vibration frequency characteristics through Fourier transform.

[0127] In one embodiment, the step of adjusting the task allocation model based on processing quality feedback includes:

[0128] A quality factor is determined, and the task allocation objective function is updated using this quality factor, where the quality factor is the processing quality score evaluated in the previous embodiment. The objective function is as follows:

[0129] ;

[0130] In the formula, It is the first The quality factor of a robotic arm task It is the total mass loss of all processed parts. These are the weighting coefficients;

[0131] Keeping the original constraints unchanged, a genetic algorithm is used to solve the updated task allocation objective function to determine the task quantity of each robotic arm in the next task unit to be processed.

[0132] In current robotic arm packaging processes, a feedback adjustment mechanism is often lacking. Uniform operating parameters are typically set to complete the entire processing. This approach cannot promptly correct for unreasonable task allocation during processing, thus failing to achieve maximum energy efficiency. Therefore, this embodiment aims to use processing task units as feedback adjustment units. After the processing of parts within the same task unit is completed, real-time processing quality data is acquired to optimize the task allocation model, ensuring a more reasonable task allocation for the next processing task unit.

[0133] Specifically, this embodiment includes the following steps:

[0134] 1) After each machining task unit is completed, use quality inspection equipment (such as vision inspection systems, force sensors, etc.) to collect packaging quality data. Based on the dimensional accuracy, surface quality, defect density, and process stability of each machined part, and their corresponding weights, calculate the machining quality score through weighted summation. As the quality factor for the processing task completed by each robotic arm, the preferred value range is 0-100, where 100 represents the highest quality and 0 represents the lowest quality.

[0135] 2) After obtaining the quality factor for each robotic arm task, calculate the total mass loss of all processed parts. .

[0136] 3) Introduce the quality factor into the original task allocation objective function to update the objective function, keeping the original constraints unchanged. Use a genetic algorithm to solve the updated task allocation objective function. The final optimized task allocation scheme can be used for the next task unit.

[0137] Therefore, this embodiment incorporates a quality factor and total quality loss into the task allocation model to ensure that high-quality tasks are prioritized. Through a dynamic feedback adjustment mechanism, the task allocation model is continuously optimized, reducing the occurrence of low-quality tasks and rework due to quality defects, thereby improving overall processing quality and reducing production costs. This allows for flexible adaptation to changes in production demands, improving the system's adaptability and robustness.

[0138] In summary, the beneficial effects that this invention can achieve are as follows:

[0139] 1) This invention calculates urgency by obtaining the order placement time, dynamically determines task priority, and divides processing task units, solving the problem that traditional systems cannot respond to changes in order urgency in real time. Dynamic priority scheduling allows urgent orders to be processed first, shortening the average order processing time; the parallel subtask processing mechanism fully utilizes the collaborative capabilities of multiple robotic arms, improving the task throughput per unit time.

[0140] 2) A task allocation model was established based on the processing system configuration parameters, such as the material arrival rate, the number of robotic arms, and the service rate. This enabled dynamic optimization of robotic arm resources, avoiding resource idleness and overload. The task allocation model based on system configuration parameters resulted in a more balanced load on the robotic arms and a significant improvement in resource utilization.

[0141] 3) By monitoring the force and trajectory of the robotic arm and making real-time adjustments using a preset algorithm, the problem of insufficient adaptability to parameter changes during processing in traditional systems is solved. The real-time monitoring and adaptive adjustment mechanism can compensate for processing errors caused by factors such as changes in material properties and mechanical wear, effectively improving processing accuracy.

[0142] 4) Immediately after the completion of each task unit, the processing quality is evaluated, and feedback is used to adjust the task allocation model, forming a closed-loop control that solves the problem of lagging quality feedback in traditional systems. This quality feedback adjustment mechanism enables the system to correct processing deviations in a timely manner, reducing the defect rate and significantly saving production costs. It also allows the multi-axis robot processing system to quickly adapt to changes in order demand and fluctuations in the production environment, providing technical support for achieving flexible manufacturing.

[0143] See Figure 4 In one embodiment, the present invention also provides a machine vision processing assistance system for a multi-axis robot, the system comprising:

[0144] Task partitioning unit 100 is used to obtain the order placement time of the task to be processed to determine the order urgency, determine the priority of the task to be processed based on the order urgency, and divide the task to be processed into multiple processing task units according to the priority order. Each processing task unit contains multiple subtasks that can be processed in parallel.

[0145] The parameter acquisition unit 200 is used to acquire the configuration parameters of the processing system, including the arrival rate of the task to be processed, the number of robotic arms of the multi-axis robot, and the service rate of each robotic arm.

[0146] The task allocation unit 300 is used to determine the task allocation model based on the configuration parameters of the processing system when executing any processing task unit; monitor the force and motion trajectory of the robotic arm of the processing task unit when the processing system is working, and adaptively adjust the force and motion trajectory of the robotic arm according to a preset algorithm;

[0147] The iterative optimization unit 400 is used to evaluate the processing quality of the current processing task unit in real time after the processing task unit is completed, adjust the task allocation model according to the processing quality feedback, and use the updated task allocation model for the next processing task unit.

[0148] It is understood that the system provided in this embodiment has functions or includes modules that can be used to execute the methods described in the above method embodiments. The specific implementation can be referred to the description of the above method embodiments, and for the sake of brevity, it will not be repeated here.

[0149] The present invention also provides an electronic device including a processor and a memory, the memory being used to store computer program code, the computer program code including computer instructions, wherein when the processor executes the computer instructions, the electronic device performs a method as described in any of the above possible implementations.

[0150] The present invention also provides a computer-readable storage medium storing a computer program, the computer program including program instructions that, when executed by a processor of an electronic device, cause the processor to perform a method as described in any of the above possible implementations.

[0151] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0152] Those skilled in the art will readily understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here. Those skilled in the art will also readily understand that the various embodiments of the present invention have different focuses, and for the sake of convenience and brevity, the same or similar parts may not be repeated in different embodiments. Therefore, parts not described or not described in detail in one embodiment can be referred to in other embodiments.

Claims

1. A machine vision-assisted processing method for multi-axis robots, characterized in that, The method includes: The order urgency is determined by the order placement time of the task to be processed. The priority of the task to be processed is determined based on the order urgency. The task to be processed is divided into multiple processing task units according to the priority order. Each processing task unit contains multiple subtasks that can be processed in parallel. Obtain the configuration parameters of the processing system, including the arrival rate of the task to be processed, the number of robotic arms of the multi-axis robot, and the service rate of each robotic arm. The service rate of each robotic arm is calculated in the following way: The basic service rate is calculated based on the total welding time of a robotic arm, tool switching time, and positioning accuracy compensation coefficient. The basic service rate is adjusted in real time based on the load attenuation coefficient, collaborative waiting loss, and fault efficiency reduction factor to obtain a dynamically updated real-time service rate. When executing any processing task unit, the task allocation model is determined according to the configuration parameters of the processing system; the force and motion trajectory of the robotic arm of the processing task unit are monitored when the processing system is working, and the force and motion trajectory of the robotic arm are adaptively adjusted according to the preset algorithm; The step of determining the task allocation model based on the configuration parameters of the processing system when executing any processing task unit includes: The urgency level is calculated based on the order placement time, delivery cycle, and task complexity coefficient. For the highest priority of urgency, a greedy algorithm is used to directly assign robotic arms, selecting the robotic arm with the highest service rate and an idle rate greater than 50%. For those with the second highest urgency priority, an improved genetic algorithm is used to assign robotic arms; After a processing task unit is completed, the processing quality of the current processing task unit is evaluated in real time. The task allocation model is adjusted based on the processing quality feedback, and the updated task allocation model is used for the next processing task unit. The real-time evaluation of the processing quality of the current processing task unit includes obtaining evaluation indicators, including the dimensional accuracy, surface quality, defect density and process stability of the processed parts, calculating the weight of the corresponding indicators, and calculating the processing quality score based on the weighted sum of each evaluation indicator.

2. The machine vision processing assistance method for multi-axis robots according to claim 1, characterized in that, The method of assigning robotic arms using an improved genetic algorithm includes: Chromosome encoding is performed, and the matching method between the task and the robotic arm is represented in binary. Determine the fitness function by calculating the weighted sum of total time and total energy consumption; Determine the mutation rules and retain the three best solutions without mutation.

3. The machine vision processing assistance method for multi-axis robots according to claim 1, characterized in that, The method further includes: The task allocation results are displayed using a two-dimensional matrix. When the actual speed of the robotic arm decreases by 10%, 30% of the tasks of the corresponding robotic arm are automatically transferred to the adjacent robotic arm.

4. The machine vision processing assistance method for multi-axis robots according to claim 1, characterized in that, The monitoring and processing system monitors the force and trajectory of the robotic arm in the processing task unit during operation, and adaptively adjusts the force and trajectory of the robotic arm according to a preset algorithm. ； ； In the formula, It is the real-time adjustable force of the robotic arm. It is the initial force of the robotic arm. It is the error between the target force and the actual force. , , These are the proportional coefficient, integral coefficient, and differential coefficient; It is the real-time adjustable movement speed of the robotic arm. It is a Jacobian matrix, representing the relationship between the joint angles of the robotic arm and the position of the end effector. It is the error between the target position and the actual position.

5. A machine vision processing assistance system for a multi-axis robot, used to implement the machine vision processing assistance method for a multi-axis robot as described in any one of claims 1-4, characterized in that, The system includes: The task division unit is used to obtain the order placement time of the task to be processed to determine the order urgency, determine the priority of the task to be processed based on the order urgency, and divide the task to be processed into multiple processing task units according to the priority order. Each processing task unit contains multiple subtasks that can be processed in parallel. The parameter acquisition unit is used to acquire the configuration parameters of the processing system. These parameters include the arrival rate of the task to be processed, the number of robotic arms in the multi-axis robot, and the service rate of each robotic arm. The service rate of each robotic arm is calculated in the following way: The basic service rate is calculated based on the total welding time of a robotic arm, tool switching time, and positioning accuracy compensation coefficient. The basic service rate is adjusted in real time based on the load attenuation coefficient, collaborative waiting loss, and fault efficiency reduction factor to obtain a dynamically updated real-time service rate. A task allocation unit is used to determine a task allocation model based on the configuration parameters of the machining system when executing any machining task unit; monitor the force and trajectory of the robotic arm of the machining task unit when the machining system is working, and adaptively adjust the force and trajectory of the robotic arm according to a preset algorithm; the step of determining the task allocation model based on the configuration parameters of the machining system when executing any machining task unit includes: The urgency level is calculated based on the order placement time, delivery cycle, and task complexity coefficient. For the highest priority of urgency, a greedy algorithm is used to directly assign robotic arms, selecting the robotic arm with the highest service rate and an idle rate greater than 50%. For those with the second highest urgency priority, an improved genetic algorithm is used to assign robotic arms; An iterative optimization unit is used to evaluate the processing quality of the current processing task unit in real time after it is completed, adjust the task allocation model according to the processing quality feedback, and apply the updated task allocation model to the next processing task unit. The real-time evaluation of the processing quality of the current processing task unit includes obtaining evaluation indicators, including the dimensional accuracy, surface quality, defect density and process stability of the processed part, calculating the weight of the corresponding indicators, and calculating the processing quality score according to the weighted sum of each evaluation indicator.

6. An electronic device, characterized in that, include: A processor and a memory, the memory being used to store computer program code, the computer program code including computer instructions, wherein when the processor executes the computer instructions, the electronic device performs the machine vision processing assistance method for a multi-axis robot as described in any one of claims 1 to 4.

7. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, the computer program including program instructions, which, when executed by a processor of an electronic device, cause the processor to perform the machine vision processing assistance method for a multi-axis robot as described in any one of claims 1 to 4.

Images