Flexible operation control system based on mechanical arm control algorithm

Abstract

The invention relates to a flexible operation control system based on a mechanical arm control algorithm, and aims at solving the problems that an existing mechanical arm is insufficient in sensing capacity, poor in operation flexibility, low in multi-task scheduling efficiency and the like in a complex environment. The system comprises a multi-mode sensor module, a control algorithm module, a flexible execution module, a dynamic task allocation and scheduling module and a cloud collaboration and learning module. The multi-mode sensor module obtains comprehensive information of the operation environment of the mechanical arm in real time through a force and torque sensor, a visual sensor, a touch sensor and the like. And the control algorithm module adopts path planning, adaptive impedance control and model prediction control to realize dynamic obstacle avoidance, flexible grabbing and high-precision operation of the mechanical arm in a complex environment. And the flexible execution module is combined with a multi-degree-of-freedom mechanical arm and a flexible clamping device. And the dynamic task allocation and scheduling module optimizes the task sequence according to the task priority and the delay time. And the cloud collaboration and learning module continuously optimizes a control algorithm and a scheduling model through reinforcement learning.

Description

Technical Field

[0001] The present invention relates to the technical field of robot control systems, and in particular to a flexible operation control system based on a robot control algorithm. Background Art

[0002] With the rapid development of industrial automation technology, robotic arms, as core equipment for intelligent manufacturing, have been widely used in assembly, handling, flexible grasping, and high-precision operation. However, existing robotic arm control systems still have obvious deficiencies in complex task scenarios. First, most robotic arms rely on a single type of sensor (such as a visual sensor or a force sensor) to obtain environmental information, but a single sensor is difficult to fully perceive the operating scenario in a dynamic and complex environment. For example, the accuracy of visual sensors decreases in high temperature and strong light environments, and the feedback of force sensors is not accurate enough under dynamic loads or multiple contact points. In addition, the data acquisition and system feedback speeds of existing sensors are slow, and path deviations or even collisions are prone to occur during high-speed operations or real-time obstacle avoidance. This limitation in perception capabilities significantly reduces the adaptability of robotic arms in flexible grasping, dynamic obstacle avoidance, and complex working conditions.

[0003] Secondly, most existing control algorithms use traditional fixed parameter methods (such as proportional integral differential control or impedance control) and lack dynamic adjustment capabilities. For example, when flexibly grasping lightweight or fragile objects, fixed impedance parameters may cause excessive grasping force and damage the objects; while in the operation of high-rigidity objects, the grasping force may not be enough to complete the task. In terms of path planning, existing technologies find it difficult to achieve real-time obstacle avoidance and dynamic path adjustment in complex scenarios. Once the obstacle or target position changes, traditional algorithms often cause mission failure due to lagging path updates. In addition, the insufficient model prediction capability makes it difficult for the system to quickly adapt to nonlinear dynamic changes, further limiting the operational flexibility of the robotic arm.

[0004] At the same time, in multi-task collaborative production lines, the existing task scheduling mechanism of robotic arms also has limitations. Current task scheduling is mostly based on static rules or fixed orders, and lacks dynamic optimization of task priorities and delay times. For example, high-priority tasks cannot be executed in a timely manner due to scheduling rules, and the delay time of low-priority tasks is not reasonably controlled. These problems will lead to a decrease in overall production efficiency. In addition, existing systems generally lack data-driven optimization capabilities, cannot use historical task data to improve scheduling models or control parameters, and are inefficient and lack adaptability when executing tasks for a long time.

[0005] In summary, the existing robotic arm control system has obvious deficiencies in multimodal perception, dynamic adaptive control, intelligent task scheduling, and cloud data optimization. These problems limit the breadth and depth of application of robotic arms in complex environments. Therefore, designing a flexible operation control system with multimodal perception, dynamic adaptive control, efficient task scheduling, and data-driven optimization capabilities has become the key to solving existing problems, in order to meet the needs of complex task scenarios and significantly improve the operational efficiency and adaptability of the robotic arm.

[0006] To this end, we urgently need to design a flexible operation control system based on the robotic arm control algorithm to solve the above problems. Summary of the invention

[0007] The present invention provides a flexible operation control system based on a mechanical arm control algorithm to solve the defects in the prior art.

[0008] Flexible operation control system based on robot arm control algorithm, including:

[0009] The multimodal sensor module is used to collect robot arm operation and environmental data in real time. The sensors and their output data are as follows:

[0010] Force and torque sensor: contact force F x ,F y ,F z and contact torque τ x ,τ y ,τ z , respectively represent the forces along the X, Y, and Z axes and the moments around the X, Y, and Z axes;

[0011] Vision sensor: The three-dimensional position of the target object (x v ,y v ,z v ) and attitude angle θ v ;

[0012] Position and attitude sensor: The position of the end of the robot arm (x now ,y now ,z now ),speed Acceleration

[0013] Torque joint sensor: torque of each joint

[0014] Tactile sensor: contact pressure P c and contact area A c ;

[0015] Temperature sensor: Temperature T at the end of the robot arm e and the target object temperature T o;

[0016] Ultrasonic or laser ranging sensor: obstacle distance d i ;

[0017] The control algorithm module takes the data from the sensor module as input and generates output based on the following algorithms: Path planning algorithm: inputs the target position and current state, and outputs the expected trajectory (x d ,y d ,z d ) and attitude angle θ d ;

[0018] Adaptive impedance control algorithm: input external force and joint torque, output end force F(t) and torque τ(t);

[0019] Model predictive control algorithm: Input the desired trajectory (x d ,y d ,z d ) and the current state, output the optimal control input u(k), including joint forces and joint torque

[0020] a flexible execution module, including a multi-degree-of-freedom manipulator and a flexible gripper, for executing operations generated by the control algorithm module;

[0021] Dynamic task allocation and scheduling module, dynamically adjusts operation strategies based on sensor feedback and task priorities;

[0022] Cloud collaboration and learning module, used to store sensor data and optimize control algorithm models.

[0023] Preferably, the path planning algorithm is based on the target position (x v ,y v ,z v ) and the current end position (x now ,y now ,z now ) is input to generate the expected trajectory:

[0024] x d =x now +η x (x v -x now ),

[0025] y d =y now +η y (y v -y now ),

[0026] z d =znow +η z (z v -z now ),

[0027] Where: x d ,y d ,z d : expected trajectory; η x ,η y ,η z : Step size ratio.

[0028] 3. The flexible operation control system based on the robot arm control algorithm according to claim 2, characterized in that the path planning algorithm uses ultrasonic or laser ranging sensor data d i As input, the following obstacle avoidance constraints are satisfied:

[0029] d i ≥d min ,

[0030] Where: d i : The distance to the ith obstacle; d min :Safe distance.

[0031] Preferably, the adaptive impedance control algorithm uses force sensor data (F x ,F y ,F z ), torque data (τ x ,τ y ,τ z ) and joint torque sensor data For input, dynamically adjust the impedance parameters:

[0032] K=K0+α·F contat ,

[0033] B=B0+β·τ contact ,

[0034]

[0035] Among them: K, B, M: stiffness, damping and inertia parameters;

[0036] External force modulus; External moment modulus; Maximum joint torque.

[0037] Preferably, the model predictive control algorithm uses the current state (x now ,y now ,z now ) and the desired trajectory generated by path planning (x d ,yd ,z d ) is input and the objective function is optimized:

[0038]

[0039] in:

[0040] x(k)=[x now ,y now ,z now ] T : Current state of the end of the robot arm;

[0041] x ref (k) = [x d ,y d ,z d ] T : expected trajectory point;

[0042] Optimize control input, including joint forces and torques; λ: regularization coefficient.

[0043] Preferably, the optimized control input u(k) satisfies the following constraints:

[0044]

[0045] in: The minimum and maximum forces allowed for the ith joint; The minimum and maximum moments allowed for the ith joint.

[0046] Preferably, the tactile sensor is provided with a contact pressure P c and contact area A c is the input used to control the gripping force of the flexible gripper, satisfying the following conditions:

[0047] F grip =P c ·A c ≤F safe ,

[0048] Among them: F grip : Grasping force; F safe : Safe grasping force threshold.

[0049] Preferably, the temperature sensor is based on the temperature T of the end of the robot arm. e and the target object temperature T o is the input for operation constraints, satisfying the following conditions:

[0050] T safe =|T e -T o |≤ΔT,

[0051] Where: ΔT: allowable temperature difference range.

[0052] Preferably, the dynamic task allocation module is based on the task time requirement T i 、Task complexity C i and task energy consumption E i As input, calculate the task priority:

[0053] P i =w1T i +w2C i +w3E i ,

[0054] Among them: w1, w2, w3: priority weights.

[0055] Preferably, the task scheduling is based on the task priority P i and task delay time D i As input, optimize the following objective function:

[0056]

[0057] Where: P i : The priority of the i-th task;

[0058] D i =T finish,i -T plan,i : The delay time of the i-th task;

[0059] T flnish,i : The actual completion time of the i-th task is recorded by the flexible execution module;

[0060] T plan,i : The planned completion time of the i-th task is pre-set by the task scheduling module;

[0061] N: Total number of tasks.

[0062] Beneficial effects: This technical solution significantly improves the perception ability of the robot arm in complex scenes by introducing a multimodal sensor module. Multimodal sensors (including force and torque sensors, visual sensors, tactile sensors, etc.) realize comprehensive perception of the robot arm's surrounding environment and operating status. For example, the visual sensor captures the target position and dynamic obstacle information in real time, the tactile sensor monitors the clamping pressure, and the temperature sensor makes the operation safe. This multi-sensor collaborative work overcomes the limitations of a single sensor, enabling the robot arm to operate accurately in dynamic and diverse environments, while reducing operational errors and improving the quality of task completion.

[0063] Through the path planning, adaptive impedance control and model predictive control functions of the control algorithm module, this technical solution significantly improves the flexible operation capability and environmental adaptability of the robot arm. The dynamic path planning algorithm realizes real-time obstacle avoidance and trajectory optimization in complex environments. The adaptive impedance control can adjust the impedance parameters of the robot arm in real time according to the changes in external forces and torques, thereby ensuring safety when grasping lightweight or fragile objects and providing sufficient operating force when carrying high-rigidity objects. In addition, model predictive control uses the optimization objective function to generate optimal control instructions, allowing the robot arm to efficiently respond to nonlinear dynamic changes and perform well in high-precision and high-dynamic demand scenarios.

[0064] In multi-task collaboration and long-term task execution scenarios, this technical solution improves the overall task scheduling efficiency and system self-optimization capabilities through dynamic task allocation and scheduling modules and cloud collaboration and learning modules. The dynamic task allocation mechanism optimizes the task sequence in real time according to task priority and delay time, so that high-priority tasks can be executed in a timely manner. The cloud collaboration module optimizes path planning, adaptive control parameters and task scheduling models through reinforcement learning, thereby achieving continuous iterative upgrades of the control system. This mechanism not only improves the long-term execution efficiency of the system, but also enhances its adaptability to new task scenarios. In summary, this technical solution has the significant advantages of comprehensive perception, flexible operation, intelligent scheduling and strong self-optimization capabilities, providing a more efficient and reliable solution for the application of robotic arms in complex environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0066] Figure 1 This is a system block diagram of a ventilation control system based on the Internet of Things. DETAILED DESCRIPTION

[0067] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0068] The following is combined with Figure 1 With reference to the accompanying drawings and a number of embodiments, specific embodiments of the present invention are described in detail.

[0069] 1: System architecture, this flexible operation control system includes the following modules:

[0070] Multimodal sensor module: collects status information of the robotic arm and the environment in real time.

[0071] Control algorithm module: including path planning, adaptive impedance control, model predictive control, etc.

[0072] Flexible execution module: includes a multi-degree-of-freedom robotic arm and a flexible gripper, responsible for executing control instructions.

[0073] Dynamic task allocation and scheduling module: optimize task sequence and reduce task delays.

[0074] Cloud collaboration and learning module: data storage and model optimization.

[0075] 2. Working principle of each module;

[0076] Multimodal sensor module: force and torque sensor, output: contact force F x ,F y ,F z and contact torque τ x ,τ y ,τ z . Application: used to calculate the external force modulus F contact and the external moment modulus τ contact :

[0077]

[0078] Vision sensor: Output: Target object position (x v ,y v ,z v ) and attitude angle θ v . Purpose: Provide target position input for path planning.

[0079] Position and attitude sensor: Output: The position of the end of the robot arm (x now ,y now ,z now ),speed Acceleration Purpose: Used for state feedback of path planning and control algorithms.

[0080] Tactile sensor: Output: Contact pressure P c and contact area A c . Purpose: Calculate the clamping force F grip :F grip =P c ·A c .

[0081] Temperature sensor: Output: Temperature T at the end of the robot arm e and the target object temperature T o Purpose: Used to determine safety constraints:

[0082] T safe =|T e -T o |≤ΔT.

[0083] Ultrasonic or laser ranging sensor: Output: obstacle distance d i . Purpose: Obstacle avoidance constraints for path planning:

[0084] d i ≥d min .

[0085] Control algorithm module: path planning algorithm, input: target position (x v ,y v ,z v ) and the current end position (x n ow,y now ,z now ). Output: expected trajectory (x d ,y d ,z d ) and attitude angle θ d .

[0086] Formula: x d =x now +η x (x v -x now ),y d =y now +η y (y v -y now ),z d =z now

[0087] where η x ,η y ,η z is the step size coefficient.

[0088] Adaptive impedance control algorithm input: external force modulus F c ontact, external moment modulus τ c ontact, joint torque

[0089] Output: end force F(t) and torque τ(t).

[0090] formula:

[0091] Among them, the impedance parameters are adjusted dynamically:

[0092] K=K0+α·F contact ,B=B0+β·τ contact .

[0093] Model predictive control algorithm: Input: current state (x n ow,y now ,z now ), expected trajectory (x d ,y d ,z d ).

[0094] Output: Optimized control input u(k). Formula:

[0095]

[0096] constraint:

[0097] Dynamic task allocation and scheduling module: task priority calculation, input: task time requirement T i Complexity C i 、Energy consumption i .

[0098] Output: Task priority P i :

[0099] P i =w1T i +w2C i +w3E i .

[0100] Task delay time calculation: Input: planned completion time T plan,i , Actual completion time T finish,i .

[0101] Scheduling optimization goal, output: delay time D i :

[0102] D i =T finish,i -T plan,i .

[0103] formula:

[0104] Cloud collaboration and learning module: Data upload: real-time upload of sensor data, task completion time, etc.

[0105] Model optimization: Optimize path planning, adaptive impedance control and task scheduling models based on reinforcement learning.

[0106] Parameter update: The optimized model parameters are sent to the local controller.

[0107] The following is combined with several specific embodiments to further illustrate:

[0108] Example 1: Background description: A robotic arm needs to carry parts from a storage area to an assembly table for precision assembly. The task requires the robotic arm to avoid excessive force that may damage the parts during the assembly process while ensuring high-precision assembly.

[0109] Specific implementation steps: The visual sensor obtains the location of the parts in the storage area (x v ,y v ,z v ) and assembly position

[0110] (x target ,y target ,z target) .

[0111] The position sensor provides the initial position of the end of the robot arm (x now ,y now ,z now ).

[0112] The system generates the expected trajectory based on the input target position and the current end position of the robot arm:

[0113] η x (x v -x now ),y d =y now +η y (y v -y now ),z d =z r

[0114] Avoid possible obstacles so that the path meets the constraint d i ≥d min .

[0115] Handling process control: The adaptive impedance control algorithm dynamically adjusts the impedance parameters based on the desired trajectory of the path planning and the real-time feedback of the force sensor:

[0116]

[0117] Among them, K, B, and M are adjusted dynamically:

[0118] K=K0+α·F contact ,B=B0+β·τ contact .

[0119] This allows the robotic arm to run stably during the handling process and avoid path deviation due to external interference.

[0120] Assembly Control: Force and Torque Sensors Monitor Assembly Contact Forces and Torques:

[0121]

[0122] According to the detection data, the assembly force at the end of the robot arm is dynamically adjusted so that the assembly torque meets the accuracy requirements.

[0123] Task completion and recording: The system records the actual completion time of the task T f inish and delay time D i =T finish,i -T plan,i .

[0124] The data is uploaded to the cloud and used to optimize the assembly model.

[0125] Example 2: The robot arm needs to grab fragile objects (such as eggs or fruits) in a complex environment, avoid dynamic obstacles, and move the objects to a designated target area.

[0126] The visual sensor captures the position of the target object (x v ,y v ,z v ) and obstacle distribution. The ultrasonic ranging sensor measures the obstacle distance d in real time i The tactile sensor initializes the gripper contact pressure P c and contact area A c .

[0127] Based on the target position and obstacle position, dynamically plan a collision-free trajectory:

[0128] x d =f(x now ,d i ), satisfying d i ≥d min .

[0129] If the obstacle distance is less than d min , replan the trajectory to bypass obstacles.

[0130] Grasping control: Tactile sensors provide real-time feedback on gripper contact pressure and contact area, dynamically adjusting gripping force:

[0131] F grip =P c ·A c ,

[0132] Make F grip ≤F safe, to avoid damage to items.

[0133] Transport task execution: The flexible execution module transports items to the target area according to the planned path.

[0134] Adaptive impedance control handles external interference during handling, allowing items to be grasped stably and carried safely.

[0135] Task completion and recording: The system records the task completion time T finish,i and transportation path data, and upload them to the cloud for optimization model.

[0136] Example 3: The robot arm needs to move heated industrial products from a high temperature area to a cooling area while avoiding contact with other high temperature objects.

[0137] The temperature sensor monitors the temperature T at the end of the robot arm e and the target object temperature T o The system determines whether the safety constraints are met:

[0138] T safe =|T e -T o |≤ΔT.

[0139] If not satisfied, the operation is suspended and the user is prompted to adjust the environment.

[0140] The system dynamically plans obstacle avoidance paths based on the distribution of high-temperature areas, making the trajectory safe:

[0141] x d =f(x now ,d i ), satisfying d i ≥d min .

[0142] If the robot arm detects a high-temperature obstacle during transportation, it will re-plan the path to bypass the obstacle. The force and torque sensor monitors the stability of the object during transportation to prevent it from slipping due to external interference.

[0143] Adaptive impedance control dynamically adjusts the motion stiffness and damping of the robot arm to adapt to different handling conditions.

[0144] Task Completion and Recording After the object is safely moved to the cooling area, the task completion time and ambient temperature data are recorded and uploaded to the cloud.

[0145] Example 4: A robotic arm performs multiple tasks on a production line (such as parts assembly, material handling, and finished product testing). The scheduling order needs to be dynamically adjusted according to the task priority to improve overall production efficiency.

[0146] System input task time requirement Y i Complexity Ci 、Energy consumption i . Calculate the priority of each task:

[0147] P i =w1T i +w2C i +w3E i .

[0148] According to the priority P i and delay time D i ,Optimize the task scheduling order:

[0149]

[0150] Adjust the scheduled time T of low priority tasks in real time plan,i , reducing overall latency.

[0151] Complete each task in sequence: use force sensors to monitor assembly force during parts assembly, dynamically avoid obstacles during material handling, and obtain inspection results by combining finished product inspection with visual sensors.

[0152] After each task is completed, record the actual completion time T finish,io

[0153] Cloud optimization: upload task data, optimize the scheduling model, and improve the execution efficiency of subsequent tasks.

[0154] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A flexible operation control system based on a robotic arm control algorithm, characterized in that: include: The multimodal sensor module is used to collect robot arm operation and environmental data in real time. The sensors and their output data are as follows: Force and torque sensor: contact force F x ,F y ,F x and contact torque τ x ,τ y ,τ z , respectively represent the forces along the X, Y, and Z axes and the moments around the X, Y, and Z axes; Vision sensor: The three-dimensional position of the target object (x v ,y v ,z v ) and attitude angle θ v ; Position and attitude sensor: The position of the end of the robot arm (x now ,y now ,z now ),speed Acceleration Torque joint sensor: torque of each joint Tactile sensor: contact pressure P c and contact area A c ; Temperature sensor: Temperature T at the end of the robot arm e and the target object temperature T o ; Ultrasonic or laser ranging sensor: obstacle distance d i ; Control algorithm module: takes the data from the sensor module as input and generates output based on the following algorithms: Path planning algorithm: inputs the target position and current state, and outputs the expected trajectory (x d ,y d ,z d ) and attitude angle θ d ; Adaptive impedance control algorithm: input external force and joint torque, output end force F(t) and torque τ(t); Model predictive control algorithm: Input the desired trajectory (x d ,y d ,z d ) and the current state, output the optimal control input u(k), including joint forces and joint torque Flexible execution module: including a multi-degree-of-freedom manipulator and a flexible gripper, used to execute the operations generated by the control algorithm module; Dynamic task allocation and scheduling module: dynamically adjusts operation strategies based on sensor feedback and task priorities; Cloud collaboration and learning module: used to store sensor data and optimize control algorithm models.

2. The flexible operation control system based on the robot arm control algorithm as claimed in claim 1, characterized in that: The path planning algorithm is based on the target position (x v ,y v ,z v ) and the current end position (x now ,y now ,z now ) is input to generate the expected trajectory: x d =x now +η x (x v -x now ), and d =and now +η y (and v -and now ), With d =from now +η z (With v -With now ), Where: x d ,y d ,z d : expected trajectory; η x ,η y ,η z : Step size ratio.

3. The flexible operation control system based on the robot arm control algorithm as claimed in claim 2, characterized in that: The path planning algorithm uses ultrasonic or laser ranging sensor data d i As input, the following obstacle avoidance constraints are satisfied: d i ≥d min , Where: d i : The distance to the ith obstacle; d min :Safe distance.

4. The flexible operation control system based on the robot arm control algorithm as claimed in claim 1, characterized in that: The adaptive impedance control algorithm uses the force sensor data (F x ,F y ,F z ), torque data (τ x ,τ y ,τ z ) and joint torque sensor data For input, dynamically adjust impedance parameters: K=K0+α·F contact , B=B0+β·τ contact , Among them: K, B, M: stiffness, damping and inertia parameters; External force modulus; External moment modulus; Maximum joint torque.

5. The flexible operation control system based on the robot arm control algorithm as claimed in claim 1, characterized in that: The model predictive control algorithm takes the current state (x now ,y now ,z now ) and the desired trajectory generated by path planning (x d ,y d ,z d ) is input and the objective function is optimized: in: x(k)=[x now ,y now ,z now ] T : Current state of the end of the robot arm; x ref (k) = [x d ,y d ,z d ] T : expected trajectory point; Optimize control input, including joint forces and torques; λ: regularization coefficient.

6. The flexible operation control system based on the robot arm control algorithm as claimed in claim 5, characterized in that: The optimized control input u(k) satisfies the following constraints: in: The minimum and maximum forces allowed for the ith joint; The minimum and maximum moments allowed for the ith joint.

7. The flexible operation control system based on the robot arm control algorithm as claimed in claim 1, characterized in that: The tactile sensor is provided with a contact pressure P c and contact area A c is the input used to control the gripping force of the flexible gripper, satisfying the following conditions: F grip =P c ·A c ≤F safe , Among them: F grip : Grasping force; F safe : Safe grasping force threshold.

8. The flexible operation control system based on the robot arm control algorithm as claimed in claim 1, characterized in that: The temperature sensor is based on the temperature T of the end of the robot arm. e and the target object temperature T o is the input for operation constraints, satisfying the following conditions: T safe =|T e -T o |≤ΔT, Where: ΔT: allowable temperature difference range.

9. The flexible operation control system based on the robot arm control algorithm as claimed in claim 1, characterized in that: The dynamic task allocation module is based on the task time requirement T i 、Task complexity C i and task energy consumption E i As input, calculate the task priority: P i =w1T i +w2C i +w3E i , Among them: w1, w2, w3: priority weights.

10. The flexible operation control system based on the robot arm control algorithm as claimed in claim 9, characterized in that: The task scheduling is based on the task priority P i and task delay time D i As input, optimize the following objective function: Where: P i : The priority of the i-th task; D i =T finish,i -T plan,i : The delay time of the i-th task; T finish,i : The actual completion time of the i-th task is recorded by the flexible execution module; T plan,i : The planned completion time of the i-th task is pre-set by the task scheduling module; N: Total number of tasks.

Images