Industrial robot control system and method

Abstract

The invention provides an industrial robot control system and method. The system comprises a trajectory planning module, a dynamic compensation module, a feedback control module, an execution module, a sensor module and a data storage module. The trajectory planning module generates an expected trajectory of an end effector of the robot; the dynamic compensation module calculates a dynamic compensation torque based on the expected trajectory and a robot dynamics model; the feedback control module generates feedback control torque according to the deviation between the real-time pose data acquired by the sensor module and the expected trajectory; the execution module drives robot joints to move according to the synthesized moment of the dynamic compensation moment and the feedback control moment; the sensor module collects robot joint angle, angular velocity and torque data in real time; the data storage module stores robot dynamic model parameters, control algorithm parameters and historical motion data. High-precision and smooth robot motion control can be achieved, the dynamic response and stability of the system are improved, and energy consumption and motion errors are reduced.

Description

Technical Field

[0001] The present invention relates to the field of robot control, and more particularly, to an industrial robot control system and method. Background Art

[0002] With the rapid development of industrial automation, industrial robots are increasingly widely used in the field of production and manufacturing, undertaking various complex tasks such as handling, welding, and assembly. At present, the motion control technology of industrial robots is mainly based on traditional kinematic and dynamic models. Traditional motion trajectory planning methods usually adopt simple linear interpolation or low-order polynomial interpolation. Although basic motion control can be achieved, the generated trajectory is not smooth enough, resulting in easy jitter during the movement of the robot, affecting the motion accuracy and efficiency. In terms of dynamic compensation, some existing technologies adopt dynamic models with fixed parameters, and do not fully consider the influence of factors such as changes in robot load and motion speed on dynamic characteristics, making the compensation effect poor and difficult to adapt to complex and changeable working scenarios. In the feedback control link, most adopt the traditional PID control algorithm, and the control parameters are fixed and cannot be dynamically adjusted according to the real-time motion state of the robot. When facing external disturbances and system uncertainties, the control performance will be significantly reduced.

[0003] At the same time, the existing industrial robot control system has deficiencies in data processing and utilization. The data collected by sensors is often only used for real-time control, lacking effective storage and analysis of historical data, and unable to accumulate experience from long-term operation to optimize system performance. Moreover, the collaborative work between different modules is not efficient enough, and there is a delay in information transmission, resulting in the response speed and stability of the entire control system being affected.

[0004] In the process of implementing the embodiments of the present invention, the inventors found that there are at least the following problems or defects in the prior art: the trajectory generated by the existing trajectory planning method has poor smoothness, resulting in jitter and low accuracy of the robot movement; the dynamic compensation does not fully consider the working condition changes, and the compensation effect is not ideal; the feedback control parameters are fixed and it is difficult to cope with external disturbances and system uncertainties; there are deficiencies in data processing and module collaborative work, affecting the system performance and response speed. Summary of the Invention

[0005] The present invention provides an industrial robot control system and method.

[0006] In the first aspect of the present invention, an industrial robot control system is provided, including:

[0007] A trajectory planning module, a dynamic compensation module, a feedback control module, an execution module, a sensor module, and a data storage module;

[0008] The trajectory planning module is used to generate the desired trajectory of the robot end effector according to the target pose;

[0009] The dynamic compensation module is used to calculate the dynamic compensation torque based on the desired trajectory and the robot dynamics model;

[0010] The feedback control module is used to generate the feedback control torque according to the deviation between the real-time pose data collected by the sensor module and the desired trajectory;

[0011] The execution module is used to drive the robot joint movement according to the combined torque of the dynamic compensation torque and the feedback control torque;

[0012] The sensor module is used to collect the robot joint angle, angular velocity and torque data in real time;

[0013] The data storage module is used to store the robot dynamics model parameters, control algorithm parameters and historical motion data;

[0014] The output end of the dynamic compensation module and the output end of the feedback control module are both connected to the input end of the execution module.

[0015] Further, the trajectory planning module includes an inverse kinematics unit and an interpolation calculation unit. The inverse kinematics unit calculates the desired angles of each joint of the robot according to the target pose. The interpolation calculation unit generates a continuous and smooth joint angle trajectory through a fifth-order polynomial interpolation algorithm, and its trajectory generation formula is:

[0016] q(t) = a0 + a1t + a2t 2 + a3t 3 + a4t 4 + a5t 5

[0017] where q(t) is the joint angle at time t, and a0 to a5 are interpolation coefficients, which are determined by the initial angle, target angle and boundary conditions.

[0018] Further, the dynamic compensation module uses the Newton-Euler equation to calculate the robot dynamics compensation torque, and its calculation formula is:

[0019] τ d = M(q)q + C(q, q)q + g(q)

[0020] where τ d is the dynamic compensation torque, M(q) is the symmetric positive definite mass matrix, C(q, q) is the Coriolis force matrix, g(q) is the gravity vector, q is the joint angle vector, q is the joint angular velocity vector, and q is the joint angular acceleration vector.

[0021] Further, the feedback control module uses a PID control algorithm to generate a feedback control torque, and its calculation formula is:

[0022]

[0023] where τ f is the feedback control torque, e is the deviation between the desired trajectory and the real-time pose, K p is the proportional coefficient, K i is the integral coefficient, and K d is the differential coefficient.

[0024] Further, the sensor module includes joint encoders and six-dimensional force sensors. The joint encoders are used to collect joint angles and angular velocities, and the six-dimensional force sensors are used to detect the contact force of the end effector.

[0025] Further, the parameters stored in the data storage module include the mass of the robot link, the position of the center of mass, the inertia tensor, the PID control coefficients, and the dynamic compensation torque threshold.

[0026] Further, the dynamic compensation module further includes an adaptive compensation unit, which online updates the dynamic parameters through the Lyapunov stability theory, and the update formula is:

[0027] Δθ = ΓY T s

[0028] where Δθ is the parameter correction amount, Γ is a positive definite adaptive gain matrix, Y is a regression matrix, and s is a linear combination of the tracking errors.

[0029] Further, the PID coefficients of the feedback control module are dynamically adjusted through fuzzy logic rules, and its fuzzy rule table includes the membership functions of the error e and the error change rate and the weighting factors of the outputs K p , K i , K d .

[0030] Further, the interpolation calculation unit of the trajectory planning module optimizes the interpolation coefficients using a genetic algorithm, and its fitness function is:

[0031] F = w1∫(q) 2 dt + w2∫(q) 2 dt

[0032] where w1 and w2 are weight coefficients, respectively representing the optimization weights for the smoothness of acceleration and energy consumption.

[0033] In the second aspect of the present invention, an industrial robot control method is provided, including the following steps:

[0034] S1. Generate the desired trajectory of the robot end effector through the trajectory planning module;

[0035] S2. Calculate the dynamic compensation torque based on the desired trajectory and the dynamic model through the dynamic compensation module;

[0036] S3. Generate the feedback control torque according to the deviation between the real-time pose and the desired trajectory through the feedback control module;

[0037] S4. Synthesize the dynamic compensation torque and the feedback control torque and then drive the robot joint movement;

[0038] S5. Real-time collect joint movement data through the sensor module and feedback it to the dynamic compensation module and the feedback control module;

[0039] S6. Update the dynamic model parameters and the control algorithm parameters according to the historical movement data.

[0040] According to the above embodiments of the present invention, it has at least the following beneficial effects: The industrial robot control system and the control method have significant advantages in many aspects. In terms of control accuracy, the trajectory planning module generates a continuous and smooth joint angle trajectory through the inverse kinematics unit and the interpolation calculation unit, and uses the fifth-order polynomial interpolation algorithm. The interpolation coefficients can also be optimized by the genetic algorithm, which can accurately plan the robot movement path; the feedback control module adopts the PID control algorithm and dynamically adjusts the PID coefficients through fuzzy logic rules, and can generate the feedback control torque in a timely and accurate manner according to the deviation between the real-time pose and the desired trajectory. The two cooperate with each other, and together with the dynamic compensation torque calculated by the dynamic compensation module based on the dynamic model, they act on the execution module to drive the robot joint movement, and can achieve high-precision motion control of the robot end effector, meeting the strict requirements of complex industrial tasks for the robot motion accuracy.

[0041] From the perspective of system performance and adaptability, the dynamic compensation module calculates the compensation torque using the Newton-Euler equation and includes an adaptive compensation unit, and updates the dynamic parameters online through the Lyapunov stability theory, which can better adapt to the changes in the dynamic characteristics of the robot under different working conditions; the joint encoder and the six-axis force sensor in the sensor module real-time collect rich movement data, and these data are fed back to the dynamic compensation module and the feedback control module to provide a basis for precise control; the data storage module stores the robot dynamic model parameters, control algorithm parameters and historical movement data, which can be conveniently called and analyzed by the system at any time, and then update the model and algorithm parameters according to the historical movement data. These designs cooperate with each other, which can improve the overall performance of the system and enhance the adaptability and robustness of the robot in a complex and changeable industrial environment. Description of the Drawings

[0042] By reading the following detailed description with reference to the accompanying drawings, the above and other objects, features, and advantages of the exemplary embodiments of the present invention will become readily understandable. In the drawings, several embodiments of the present invention are shown by way of illustration and not limitation, wherein:

[0043] Figure 1 Schematic diagram of the structure of an industrial robot control system provided in an embodiment of the present invention;

[0044] Figure 2 Schematic flow diagram of an industrial robot control method provided in an embodiment of the present invention;

[0045] Figure 3 Schematically shows a schematic diagram of the structure of an electronic device according to an embodiment of the present invention. Detailed implementation manners

[0046] The principles and spirit of the present invention will be described below with reference to several exemplary embodiments. It should be understood that these embodiments are given only to enable those skilled in the art to better understand and then implement the present invention, and are not intended to limit the scope of the present invention in any way. On the contrary, these embodiments are provided to make the present invention more thorough and complete, and to be able to convey the scope of the present invention fully to those skilled in the art.

[0047] Those skilled in the art know that the embodiments of the present invention can be implemented as a system, device, equipment, method, or computer program product. Therefore, the present invention can be specifically implemented in the following forms: completely hardware, completely software (including firmware, resident software, microcode, etc.), or a combination of hardware and software.

[0048] It should be noted that any number of elements in the drawings is for illustration and not limitation, and any naming is only for distinction and does not have any limiting meaning.

[0049] Below with reference to Figure 1 , Figure 1 Schematic diagram of the structure of an industrial robot control system provided in an embodiment of the present invention. As Figure 1 shown, an industrial robot control system 100 includes:

[0050] A trajectory planning module 101, a dynamic compensation module 102, a feedback control module 103, an execution module 104, a sensor module 105, and a data storage module 106;

[0051] The trajectory planning module is used to generate a desired trajectory of the robot end effector according to the target pose;

[0052] The dynamic compensation module is used to calculate the dynamic compensation torque based on the desired trajectory and the robot dynamics model;

[0053] The feedback control module is used to generate a feedback control torque according to the deviation between the real-time pose data collected by the sensor module and the desired trajectory;

[0054] The execution module is used to drive the robot joints to move according to the combined torque of the dynamic compensation torque and the feedback control torque;

[0055] The sensor module is used to collect the robot joint angle, angular velocity and torque data in real time;

[0056] The data storage module is used to store the robot dynamics model parameters, control algorithm parameters and historical motion data;

[0057] The output end of the dynamic compensation module and the output end of the feedback control module are both connected to the input end of the execution module.

[0058] It should be noted that the industrial robot control system involved here contains multiple important modules. The trajectory planning module is the core part of the motion planning of the whole system, and it generates a desired motion trajectory based on the target pose of the robot. The target pose refers to the specific position and posture that the robot end effector needs to reach. For example, in a three-dimensional space, the position can be represented by coordinates (x, y, z), and the posture can be described by parameters such as rotation angles. The dynamic compensation module combines the desired trajectory and the robot dynamics model to calculate the dynamic compensation torque. The dynamics model is a mathematical model that describes the relationship between physical quantities such as force, mass, and acceleration during the motion of the robot. It can help us understand the additional torque required by the robot in different motion states to overcome the influence of factors such as inertia and friction. The feedback control module generates a feedback control torque according to the deviation between the real-time pose of the robot and the desired trajectory, so as to precisely adjust the motion of the robot. The execution module drives the joints of the robot to move according to the combined control torque, so that the robot can move along the planned trajectory. The sensor module is responsible for collecting the pose data and force data of the robot, providing necessary information for feedback control and dynamic compensation. The data storage module is used to store the robot dynamics model parameters, control algorithm parameters, and historical motion data, etc. These data are very important for the optimization and adjustment of the system.

[0059] Specifically, the inverse kinematics unit in the trajectory planning module converts the target pose into the angle values of each joint of the robot through a series of mathematical calculations. Here, the common D-H (Denavit-Hartenberg) parameter method can be used to establish the kinematics model of the robot, and then the inverse solution calculation is carried out using this model. The interpolation calculation unit uses the fifth-order polynomial interpolation algorithm to generate a continuous and smooth joint angle trajectory. The general form of the fifth-order polynomial is q(t) = a0 + a1t + a2t2 +a3t 3 +a4t 4 +a5t 5 , where q(t) represents the change of joint angle with time t, and a0 - a5 are interpolation coefficients. To optimize the interpolation coefficients, a genetic algorithm can be used. By setting appropriate fitness functions, such as minimizing the motion time, minimizing the energy consumption, etc., to search for the optimal coefficient combination. The Newton - Euler equation in the dynamic compensation module is a commonly used dynamic modeling method. It is based on Newton's second law and the concept of Euler angles. By analyzing the forces and motions of each link of the robot, the required compensation torque can be calculated. The adaptive compensation unit can update the parameters in the dynamic model online according to the Lyapunov stability theory to adapt to changes in factors such as the robot load and motion speed. The feedback control module adopts the PID (Proportional - Integral - Derivative) control algorithm, and its control parameters K p (proportional coefficient), K i (integral coefficient) and K d (derivative coefficient) can be dynamically adjusted through fuzzy logic rules. Fuzzy logic rules are a series of conditional statements formulated based on experience and experimental data. For example, when the deviation is large, increase the value of K p to speed up the response speed; when the deviation persists, increase the value of K i to eliminate the steady - state error.

[0060] Preferably, in the trajectory planning module, in addition to the fifth - order polynomial interpolation algorithm, a spline interpolation algorithm can also be used, such as cubic spline interpolation, which can also generate a smooth trajectory and is more computationally efficient in some cases. For the application of the genetic algorithm, an adaptive genetic algorithm can be adopted, which can automatically adjust the crossover probability and mutation probability according to the fitness values during the evolution process, improving the search efficiency and convergence speed of the algorithm. In the dynamic compensation module, in addition to the Newton - Euler equation, the Lagrangian equation can also be used to establish the dynamic model. The Lagrangian equation starts from the perspective of energy and is more convenient for modeling some complex robot systems. In the feedback control module, in addition to the PID control algorithm, a fuzzy PID control algorithm or a neural network PID control algorithm can also be adopted, and these algorithms can further improve the control accuracy and adaptability. In the sensor module, in addition to joint encoders and six - dimensional force sensors, a vision sensor, such as a camera, can also be added, which can provide image information of the robot's surrounding environment to help the robot better perform path planning and target recognition.

[0061] In some embodiments, the trajectory planning module includes an inverse kinematics unit and an interpolation calculation unit. The inverse kinematics unit calculates the desired angles of each joint of the robot according to the target pose, and the interpolation calculation unit generates a continuous and smooth joint angle trajectory through a quintic polynomial interpolation algorithm. The trajectory generation formula is:

[0062] q(t) = a0 + a1t + a2t 2 + a3t 3 + a4t 4 + a5t 5

[0063] where q(t) is the joint angle at time t, and a0 to a5 are interpolation coefficients determined by the initial angle, target angle, and boundary conditions.

[0064] It should be noted that the inverse kinematics unit plays a key bridging role in the entire motion control system of the industrial robot. It receives the target pose information of the robot's end effector set externally. The target pose refers to the specific position and orientation of the robot's end in three-dimensional space. The position is usually accurately described by three-dimensional coordinates (x, y, z), and the orientation can be expressed by means of Euler angles, quaternions, etc. Then, the inverse kinematics unit performs inverse solution calculations based on the kinematic model of the robot. The kinematic model is a model that describes the mathematical relationship between the movements of each joint of the robot and the pose of the end effector. It can decompose the overall movement of the robot into the specific movements of each joint. Through this inverse solution calculation, the target pose is converted into the target angle values corresponding to each joint of the robot. These target angle values are like precise instructions given to each joint, guiding the joints to move to the appropriate positions, so that the robot's end can accurately reach the target pose.

[0065] Specifically, the D-H (Denavit-Hartenberg) parameter method is often used to construct the kinematic model. This method accurately describes the relative positions and movement relationships between each link of the robot by defining a series of link parameters, including link length, link twist angle, joint offset, and joint angle, etc. When using the D-H parameter method, these parameter values need to be accurately set according to the actual mechanical structure of the robot. The inverse solution calculation process often involves solving complex mathematical equations and may obtain multiple solutions. In order to screen out the valid solutions that meet the actual motion range and constraint conditions of the robot, the value range of the joint angles needs to be set. For example, some joints may only be able to move within a specific angle range. Beyond this range, the robot may not be able to operate normally or may encounter dangers such as collisions. At the same time, combined with the robot's work tasks and motion planning, the solution that makes the robot's motion path the shortest, energy consumption the lowest, or motion the smoothest can be preferentially selected.

[0066] Preferably, in addition to the D-H parameter method, the Modified D-H parameter method can also be used to construct the kinematic model. The Modified D-H parameter method can more concisely describe the motion relationship of the robot in some cases, especially suitable for some robots with special structures. In the inverse solution calculation, in addition to traditional numerical methods such as the Newton-Raphson method, symbolic methods can also be used. Symbolic methods can directly give analytical expressions, avoiding error accumulation in the numerical calculation process, and having advantages in scenarios with high calculation accuracy requirements. For the process of screening valid solutions, machine learning algorithms such as neural networks can be introduced. By training the neural network with a large amount of sample data, let it learn the optimal solution characteristics of the robot under different tasks and working conditions, so as to be able to screen out suitable solutions more quickly and accurately, improving the efficiency and accuracy of robot motion planning.

[0067] In some embodiments, the dynamic compensation module uses the Newton-Euler equation to calculate the dynamic compensation torque of the robot, and its calculation formula is:

[0068] τ d =M(q)q + C(q,q)q + g(q)

[0069] where, τ d is the dynamic compensation torque, M(q) is a symmetric positive definite mass matrix, C(q,q) is the Coriolis force matrix, g(q) is the gravity vector, q is the joint angle vector, q is the joint angular velocity vector, and q is the joint angular acceleration vector.

[0070] It should be noted that the interpolation calculation unit plays an important role in making the motion trajectory of the industrial robot smoothly transition in motion control. It uses the joint target angle values output by the inverse kinematics unit as basic data, and these target angle values are the specific angles that each joint of the robot needs to rotate to in order to reach the target pose of the end effector. A fifth-order polynomial interpolation algorithm is used to generate a continuous and smooth joint angle trajectory. Fifth-order polynomial interpolation is a mathematical method that generates a fifth-order polynomial function by fitting a series of data points, so that the robot joint angles can smoothly change according to the law described by this polynomial function between the starting point and the target point, avoiding sudden changes in joint motion and ensuring the smoothness and fluency of the robot motion. At the same time, in order to optimize the interpolation coefficients, a genetic algorithm is also used. The genetic algorithm is an optimization algorithm that simulates the biological evolution process. It continuously iteratively searches for the optimal interpolation coefficients by simulating natural selection and genetic mechanisms, so that the generated trajectory can meet some optimization goals such as the shortest motion time and the minimum energy consumption while meeting the motion requirements.

[0071] Specifically, the general form of the fifth-order polynomial used in the fifth-order polynomial interpolation algorithm is q(t) = a0 + a1t + a2t2 +a3t 3 +a4t 4 +a5t 5 , where q(t) represents the change of the joint angle with time t, and a0 - a5 are the interpolation coefficients to be determined. To determine these coefficients, it is necessary to construct and solve a system of equations based on the boundary conditions such as the starting angle, target angle, velocity, and acceleration at the starting and target times of the joint. For example, assuming that at the starting time t0, the joint angle is q0, the velocity is v0, and the acceleration is a0; at the target time t1, the joint angle is q1, the velocity is v1, and the acceleration is a1. Substituting these conditions into the fifth-degree polynomial and its first- and second-order derivative equations can obtain a system of equations containing six equations, thereby solving the values of a0 - a5. When applying the genetic algorithm, the fitness function needs to be defined first. The fitness function is the criterion for measuring the quality of each individual (i.e., a set of interpolation coefficients). Common fitness functions can be the reciprocal of the movement time, the negative value of the energy consumption, etc., which are specifically determined according to the optimization goal. Then, through genetic operations such as selection, crossover, and mutation, the population is continuously iteratively updated until the optimal combination of interpolation coefficients is found.

[0072] Preferably, in addition to the fifth-degree polynomial interpolation algorithm, the cubic spline interpolation algorithm can also be used. The cubic spline interpolation can also generate a smooth trajectory and is relatively simple in calculation, which is more suitable for some scenarios with high requirements for computing resources or not particularly strict requirements for the smoothness of the movement trajectory. When using the genetic algorithm to optimize the interpolation coefficients, the adaptive genetic algorithm can be used. The adaptive genetic algorithm can automatically adjust the crossover probability and mutation probability according to the fitness distribution of the population during the evolution process, avoid the algorithm from converging to the local optimal solution prematurely, and improve the search efficiency and global optimization ability of the algorithm. In addition, for the design of the fitness function, multiple optimization goals can be comprehensively considered. For example, the movement time and energy consumption can be incorporated into the fitness function at the same time, and the importance of different goals can be balanced by weighted summation, so that the generated trajectory can achieve better performance in multiple aspects.

[0073] In some embodiments, the feedback control module uses the PID control algorithm to generate the feedback control torque, and its calculation formula is:

[0074]

[0075] where, τ f is the feedback control torque, e is the deviation between the desired trajectory and the real-time pose, K p is the proportional coefficient, K i is the integral coefficient, K d is the differential coefficient.

[0076] It should be noted that the dynamic compensation module plays a key role in the precise control of industrial robots. It combines the desired trajectory and the dynamic model to calculate the dynamic compensation torque. The desired trajectory is generated by the trajectory planning module, which specifies the positions and postures that the robot should reach at each moment within a certain period of time, reflecting the ideal motion path of the robot. The dynamic model is a mathematical model that describes the relationship between physical quantities such as force, mass, and acceleration during the movement of the robot. It takes into account factors such as the mass, moment of inertia, and joint friction of each link of the robot. By substituting the desired trajectory into the dynamic model for calculation, the additional torque required to overcome the influence of various physical factors under ideal motion conditions, that is, the dynamic compensation torque, can be obtained. This compensation torque can offset the interference generated by factors such as inertia, gravity, and friction during the movement of the robot, enabling the robot to move more accurately along the desired trajectory. The adaptive compensation unit is an important part of the dynamic compensation module. It uses the Lyapunov stability theory to update the parameters of the dynamic model online. The Lyapunov stability theory is a mathematical method for analyzing the stability of a system. By constructing a suitable Lyapunov function, it is judged whether the system is stable, and the parameters of the dynamic model are adjusted according to the state error of the system to improve the accuracy and adaptability of the model.

[0077] Specifically, the Newton-Euler equations are usually used to establish the dynamic model. The Newton-Euler equations are based on Newton's second law and the concept of Euler angles. By performing force analysis and motion analysis on each link of the robot, the relationship between force and acceleration is established. When establishing the model, parameters such as the mass, center-of-mass position, and moment of inertia of each link of the robot need to be accurately measured. For joint friction, the Coulomb friction model or the viscous friction model can be used for description. When the adaptive compensation unit applies the Lyapunov stability theory, it is first necessary to construct a suitable Lyapunov function. The Lyapunov function is usually a positive definite function about the system state error and parameter error. For example, the sum of the squares of the state error plus the sum of the squares of the parameter error can be selected as the Lyapunov function. Then, according to the condition that the derivative of the Lyapunov function is less than zero, the parameter update law is derived. The parameter update law specifies how to adjust the parameters of the dynamic model in real time according to the system state error to ensure the stability and tracking performance of the system.

[0078] Preferably, in addition to the Newton-Euler equations, the Lagrange equations can also be used to establish the dynamic model. The Lagrange equations start from the perspective of energy and establish the motion equations of the system by analyzing the kinetic energy and potential energy of the system. For some robotic systems with complex structures and constraints, the Lagrange equations may be more applicable. In the adaptive compensation unit, in addition to the Lyapunov stability theory, the model reference adaptive control (MRAC) method can also be adopted. The MRAC method introduces a reference model to make the output of the actual system as close as possible to the output of the reference model. According to the error between the system output and the reference model output, the parameters of the dynamic model are adjusted in real time. In addition, in order to improve the accuracy and robustness of parameter update, estimation methods such as Kalman filtering can be combined to estimate the state and parameters of the system more precisely. When measuring the parameters of each link of the robot, advanced sensor technologies and calibration methods can be adopted to improve the accuracy of parameter measurement.

[0079] In some embodiments, the sensor module includes a joint encoder and a six-axis force sensor. The joint encoder is used to collect joint angles and angular velocities, and the six-axis force sensor is used to detect the contact force of the end effector.

[0080] It should be noted that the feedback control module is crucial for the precise motion control of industrial robots. It generates a feedback control torque based on the deviation between the real-time pose of the robot and the desired trajectory. The real-time pose refers to the actual position and orientation of the robot at the current moment, which can be obtained through the sensor module. The desired trajectory is pre-planned and represents the ideal motion path of the robot. When the actual motion of the robot deviates from the desired trajectory, a deviation will occur. The feedback control module calculates the required feedback control torque based on this deviation to adjust the motion of the robot and make it return to the desired trajectory. This module adopts the PID control algorithm. PID stands for Proportional-Integral-Derivative control, which combines three control actions: proportional, integral, and derivative. Proportional control outputs the control quantity proportionally according to the magnitude of the deviation and can quickly respond to the deviation. Integral control is used to eliminate the steady-state error of the system and continuously adjusts the control quantity by integrating the deviation. Derivative control predicts the development trend of the deviation in advance according to the change rate of the deviation and enhances the stability of the system. Fuzzy logic rules are used to dynamically adjust the parameters of the PID control algorithm. Fuzzy logic is a method for dealing with uncertainty and fuzzy information. It formulates a series of rules based on experience and knowledge and dynamically changes the proportional coefficient K p of the PID, the integral coefficient K i and the derivative coefficient K d to improve the control effect.

[0081] Specifically, the calculation formula for the control quantity u(t) of the PID control algorithm is

[0082]

[0083] Among them, e(t) is the deviation at the current moment, and K p , K i , K d are the proportional, integral, and differential coefficients respectively. In the initial stage, K p , K i , K d can be initially set by methods such as the trial-and-error method and the Ziegler-Nichols method. The trial-and-error method is to continuously try different parameter values and observe the response of the system until a relatively suitable set of parameters is found; the Ziegler-Nichols method determines the parameters based on the step response characteristics of the system. Fuzzy logic rules are usually presented in the form of a fuzzy rule table. The input variables of the rule table are the deviation e and the deviation change rate Δe, and the output variables are the adjustment amounts ΔK p , ΔK i , ΔK d of the PID parameters. For example, when the deviation is large and the deviation change rate is also large, increase the value of K p to accelerate the response speed; when the deviation persists, increase the value of K i to eliminate the steady-state error. The formulation of fuzzy rules needs to combine actual control experience and system characteristics.

[0084] Preferably, in addition to the traditional PID control algorithm, improved algorithms such as the fuzzy PID control algorithm and the neural network PID control algorithm can also be adopted. The fuzzy PID control algorithm deeply integrates fuzzy logic and PID control and can adjust the PID parameters more flexibly; the neural network PID control algorithm utilizes the self-learning and adaptive capabilities of the neural network to automatically optimize the PID parameters. For the formulation of fuzzy logic rules, intelligent optimization algorithms such as the genetic algorithm and the particle swarm algorithm can be used for optimization. These algorithms can search for the optimal fuzzy rules among a large number of rule combinations to improve the control accuracy and adaptability. When determining the initial PID parameters, in addition to the trial-and-error method and the Ziegler-Nichols method, a model-based parameter tuning method can also be adopted to calculate the appropriate parameter values according to the dynamic model of the robot. In addition, in order to improve the anti-interference ability of the system, a method combining feedforward control and feedback control can be introduced to compensate for known disturbances in advance.

[0085] In some embodiments, the parameters stored in the data storage module include the mass of the robot link, the position of the center of mass, the inertia tensor, the PID control coefficient, and the dynamic compensation torque threshold.

[0086] It should be noted that the execution module is the key link for the industrial robot to achieve actual motion. It receives the dynamic compensation torque calculated by the dynamic compensation module and the feedback control torque generated by the feedback control module. These two torques are respectively used to cancel the dynamic interference in the robot's motion and correct the motion deviation. After synthesizing these two torques, a comprehensive control torque is obtained. The execution module controls the motion of the robot joints through the drive device according to this synthesized control torque. The drive device is generally a power device such as a motor, which converts the electrical signal into mechanical motion, thereby driving the robot joints to rotate, enabling the robot to move along the pre-planned trajectory and achieve various industrial tasks.

[0087] Specifically, during the torque synthesis process, usually the dynamic compensation torque and the feedback control torque are directly added. However, in practical applications, weighted processing may be required according to different situations to better balance the effects of the two torques. For example, if the dynamic interference is large during the robot's motion, the weight of the dynamic compensation torque can be appropriately increased. The selection of the drive device is determined according to factors such as the type of the robot, the load requirement, and the motion accuracy. For small and light-load robots, stepper motors can be selected, which have the advantages of simple control and low cost; for large and high-precision robots, servo motors are required, and servo motors can provide precise position and speed control. When controlling the drive device, appropriate control parameters need to be set, such as the voltage, current, and rotation speed of the motor. The setting of these parameters is determined according to the magnitude of the synthesized control torque and the motion requirements of the robot joints. For example, when the control torque is large, the voltage and current of the motor need to be increased to provide sufficient power.

[0088] Preferably, in terms of torque synthesis, in addition to simple addition or weighted addition, intelligent algorithms can also be used for optimized synthesis. For example, using the fuzzy control algorithm to dynamically adjust the synthesis ratio of the dynamic compensation torque and the feedback control torque according to the real-time motion state and environmental information of the robot, making the robot's motion more stable and precise. For the drive device, in addition to stepper motors and servo motors, hydraulic drive devices or pneumatic drive devices can also be considered. Hydraulic drive devices have the advantages of large output force and fast response speed, and are suitable for robots with heavy loads and high-speed motions; pneumatic drive devices have the characteristics of low cost and clean and pollution-free, and are suitable for some occasions with high environmental requirements. In the parameter setting of controlling the drive device, an adaptive control algorithm can be introduced to adjust the control parameters in real time according to the actual motion situation of the robot, improving the motion performance and adaptability of the robot.

[0089] In some embodiments, the dynamic compensation module further includes an adaptive compensation unit, which updates the dynamic parameters online through the Lyapunov stability theory, and the update formula is:

[0090] Δθ=ΓYT s

[0091] Among them, Δθ is the parameter correction amount, Γ is a positive definite adaptive gain matrix, Y is a regression matrix, and s is a linear combination of tracking errors.

[0092] It should be noted that the sensor module plays an important role in information acquisition in the industrial robot system. It includes two key sensors, namely the joint encoder and the six-axis force sensor. The joint encoder is used to collect the angle data of the robot joints. The joint angle is an important parameter describing the rotational position of each joint of the robot. By accurately obtaining the joint angle information, the posture of the robot can be understood in real time. The six-axis force sensor is responsible for collecting the force and torque data received by the end effector of the robot. The force and torque reflect the force conditions of the robot during the interaction with the external environment. For example, the force borne when grasping and handling objects. These collected pose data (the pose of the robot can be calculated from the joint angle data) and force data will be transmitted to the dynamic compensation module and the feedback control module in real time. The dynamic compensation module uses these data to calculate the dynamic compensation torque more accurately to adapt to various mechanical changes in the actual movement of the robot; the feedback control module judges the deviation between the actual movement of the robot and the desired trajectory based on these data, so as to generate an accurate feedback control torque.

[0093] Specifically, the joint encoder is divided into an incremental encoder and an absolute encoder. The incremental encoder determines the change in the rotation angle of the joint by detecting pulse signals. It requires an initial position as a reference and may need to be zeroed each time it is powered on. The absolute encoder can directly output the absolute angle value of the joint without the need for zeroing operation, and has higher accuracy and reliability. When selecting a joint encoder, it is necessary to determine according to the accuracy requirements and cost budget of the robot. Generally speaking, for robots with high accuracy requirements, an absolute encoder with a higher resolution can be selected. The six-axis force sensor can measure the force in three directions (usually the X, Y, and Z axis directions) and the torque in three directions (the torque around the X, Y, and Z axes) at the same time. When installing the six-axis force sensor, it is necessary to ensure that its installation position is accurate and its connection with the end effector of the robot is firm to ensure the accuracy of the measurement data. At the same time, the sensor needs to be calibrated. The calibration process usually includes zero calibration and sensitivity calibration to eliminate the errors of the sensor itself.

[0094] Preferably, in addition to the joint encoder and the six-axis force / torque sensor, other types of sensors such as vision sensors and lidar can also be added to the sensor module. The vision sensor can provide image information of the robot's surrounding environment to assist the robot in target recognition, positioning, and path planning; the lidar can obtain real-time three-dimensional spatial information around the robot to improve the robot's environmental perception ability. For the joint encoder, if the robot has extremely high requirements for the real-time performance of angle measurement, an encoder with high-speed response can be used. In terms of the calibration of the six-axis force / torque sensor, in addition to traditional calibration methods, a calibration method based on machine learning can also be adopted. By training the model with a large amount of experimental data, the calibration accuracy and efficiency can be improved. In addition, to improve the reliability of sensor data, redundant design can be adopted, that is, multiple sensors of the same type are installed to compare and fuse the collected data.

[0095] In some embodiments, the PID coefficients of the feedback control module are dynamically adjusted by fuzzy logic rules, and its fuzzy rule table includes the membership functions of the error e and the error change rate and the output K p 、K i 、K d weighting factors.

[0096] It should be noted that the data storage module is an important part of the industrial robot system for data management and storage. It mainly stores the dynamic model parameters of the robot, control algorithm parameters, and historical motion data. The dynamic model parameters describe the relationships between physical quantities such as force, mass, and acceleration during the robot's motion, and are the basis for accurate calculations in the dynamic compensation module, such as the mass, center of mass position, and moment of inertia of each link. The control algorithm parameters are the key to the effective control of modules such as the feedback control module, such as the proportional coefficient, integral coefficient, and differential coefficient in the PID control algorithm. The historical motion data records information such as the robot's past motion trajectories, joint angles, and force conditions. These data can provide a basis for system performance evaluation, fault diagnosis, and algorithm optimization. The purpose of storing these data is to facilitate the ready access of each module. For example, the dynamic compensation module can refer to the dynamic model parameters when calculating the compensation torque, and the feedback control module can draw on the historical motion data and control algorithm parameters when adjusting the control strategy.

[0097] Specifically, the storage of the kinetic model parameters needs to ensure the accuracy and integrity of the data. Before storage, these parameters need to be accurately measured and calibrated. For example, for the mass of the connecting rod, a high-precision weighing device can be used for measurement; for the moment of inertia, it can be obtained through experimental measurement or theoretical calculation. The storage format can adopt common database formats, such as relational databases (MySQL, SQL Server, etc.), and different types of parameters are stored in different tables respectively for convenient management and query. The setting of the control algorithm parameters usually needs to be adjusted according to the specific tasks and performance requirements of the robot. In the initial stage, a set of basic parameter values can be determined through the trial-and-error method or an experience-based method, and then optimized according to the motion effect of the robot during actual operation. These parameters can also be stored in the database, and the time and reason for each adjustment are recorded for subsequent analysis. The collection of historical motion data can be obtained in real time through the sensor module and then stored at regular time intervals. When storing, a combination of file storage and database storage can be adopted. For some short-term data with high real-time requirements, it can be first stored in a file and imported into the database regularly for long-term preservation.

[0098] Preferably, in terms of data storage, a distributed storage system such as the Hadoop Distributed File System (HDFS) can be adopted. The distributed storage system has advantages such as high reliability, scalability, and fault tolerance, and can better meet the storage requirements of large-scale data. For the update of the kinetic model parameters, an online update method can be adopted. According to the real-time data collected by the sensor and the operating state of the system, the model parameters are dynamically adjusted to improve the accuracy of the model. In terms of the optimization of the control algorithm parameters, artificial intelligence algorithms such as genetic algorithms and particle swarm algorithms can be introduced to automatically search for the optimal parameter combination. For the analysis of historical motion data, data mining and machine learning technologies can be adopted to mine the hidden laws and patterns in the data, providing more valuable information for the performance improvement and fault prediction of the robot. In addition, to ensure data security, encryption technology can be used to encrypt the stored data to prevent data leakage and malicious tampering.

[0099] In some embodiments, the interpolation calculation unit of the trajectory planning module optimizes the interpolation coefficient using a genetic algorithm, and its fitness function is:

[0100] F = w1∫(q) 2 dt + w2∫(q) 2 dt

[0101] where w1 and w2 are weight coefficients, respectively representing the optimization weights for the smoothness of acceleration and energy consumption.

[0102] It should be noted that this control method is a complete process for realizing the precise motion control of industrial robots. First, a desired trajectory is generated based on the target pose. The target pose refers to the specific position and orientation that the end effector of the robot needs to reach, and the desired trajectory is the motion path planned for the robot to move from the current state to the target pose. Then, the dynamic compensation torque is calculated by combining the desired trajectory and the dynamic model. The dynamic model describes the relationship between physical quantities such as force, mass, and acceleration during the motion of the robot, and the dynamic compensation torque is used to cancel the interference generated by factors such as inertia and gravity during the robot's motion. At the same time, the feedback control torque is generated according to the deviation between the real-time pose of the robot and the desired trajectory. The real-time pose is the current actual position and orientation of the robot, and the feedback control torque can correct the deviation of the robot's motion. After synthesizing the dynamic compensation torque and the feedback control torque, the robot joints are driven to move, enabling the robot to move towards the target pose. During this process, the sensor module collects the pose data and force data of the robot in real time, and these data are used for subsequent control adjustment. Finally, the parameters of the dynamic model and the control algorithm are updated based on the collected data and historical motion data to improve the control accuracy and adaptability of the system.

[0103] Specifically, when generating the desired trajectory, the inverse kinematics unit uses the D-H (Denavit-Hartenberg) parameter method to convert the target pose into joint target angle values, and the interpolation calculation unit uses the fifth-order polynomial interpolation algorithm to generate a smooth joint angle trajectory, whose general form is q(t) = a0 + a1t + a2t 2 + a3t 3 + a4t 4 + a5t 5 , and the interpolation coefficients a0 - a5 are solved by setting boundary conditions, and genetic algorithms can also be used to optimize these coefficients. When calculating the dynamic compensation torque, a dynamic model is established based on the Newton-Euler equation, considering parameters such as link mass, center of mass position, and moment of inertia, and the adaptive compensation unit updates the model parameters online according to the Lyapunov stability theory. The PID control algorithm is used to generate the feedback control torque, and the formula is

[0104]

[0105] where e(t) is the deviation, and K p 、K i 、K d are parameters, and these parameters are dynamically adjusted according to fuzzy logic rules. In the sensor module, the joint encoder collects joint angle data, and the six-axis force sensor collects the force and torque data of the end effector. When updating the parameters, appropriate algorithms (such as the least squares method, etc.) are used to optimize the parameters of the dynamic model and the control algorithm according to the collected data and historical motion data.

[0106] Preferably, in generating the desired trajectory, in addition to the quintic polynomial interpolation algorithm, a spline interpolation algorithm such as cubic spline interpolation can also be selected, and its calculation efficiency may be higher. For the dynamic model, in addition to the Newton-Euler equation, the Lagrangian equation is also an option, which models from the energy perspective and is applicable to robots with complex structures. In feedback control, in addition to the PID control algorithm, a fuzzy PID control algorithm or a neural network PID control algorithm can be adopted to improve the intelligence and adaptability of the control. When updating parameters, in addition to the least squares method, the Kalman filter algorithm can also be used, which can estimate parameters more accurately. In addition, for the sensor module, devices such as vision sensors and lidar can be added to enrich the types of collected data, improve the robot's perception ability of the environment, and further optimize the control process.

[0107] The above-mentioned various embodiments of the present invention have the following beneficial effects: In complex industrial production scenarios, the industrial robot control system described in the present invention exhibits excellent performance. The trajectory planning module can accurately generate the desired trajectory, and its interpolation calculation unit is optimized by the quintic polynomial interpolation algorithm and the genetic algorithm to ensure the smoothness and accuracy of the motion path, making the robot move smoothly and position accurately during operation; the dynamic compensation module calculates the compensation torque through the Newton-Euler equation and combines the adaptive compensation unit to update the dynamic parameters online, effectively canceling the dynamic interference during the robot's movement; the feedback control module dynamically adjusts the coefficients by means of the PID control algorithm and the fuzzy logic rules, and generates the feedback control torque in real time according to the pose deviation. With the cooperation of multiple modules, the execution module drives the joint movement according to the combined torque, which can greatly improve the motion control accuracy of the robot, ensure that the robot can accurately reach the target pose when performing complex tasks, reduce errors, and improve the product processing quality.

[0108] In addition, the close cooperation of the various modules of the system can further enhance the overall efficiency of the system. The joint encoder and six-dimensional force sensor in the sensor module collect rich motion data in real time, providing key basis for control decisions; the data storage module stores a large amount of parameters and historical motion data, which is not only convenient for each module to call at any time, but also supports updating the model and algorithm parameters based on historical data. In this way, the system can continuously optimize its own performance according to the actual working conditions, has stronger adaptability, can quickly adapt to different working environments and task requirements, effectively improve the working efficiency of industrial robots, reduce maintenance costs, and provide strong support for the automation and intelligent development of industrial production.

[0109] As Figure 2 shown, an industrial robot control method 200 of some embodiments, the method 200 includes the following steps:

[0110] S1. Generate the desired trajectory of the robot end effector through the trajectory planning module;

[0111] S2. Calculate the dynamic compensation torque based on the desired trajectory and the dynamic model by means of the dynamic compensation module;

[0112] S3. Generate a feedback control torque according to the deviation between the real-time pose and the desired trajectory by means of the feedback control module;

[0113] S4. Synthesize the dynamic compensation torque and the feedback control torque and then drive the robot joints to move;

[0114] S5. Collect the joint motion data in real time through the sensor module and feedback it to the dynamic compensation module and the feedback control module;

[0115] S6. Update the parameters of the dynamic model and the control algorithm parameters according to the historical motion data.

[0116] It can be understood that the steps described in the industrial robot control method 200 correspond to the respective modules in the industrial robot control system described in the reference Figure 1 Therefore, the modules, features and beneficial effects described above for the industrial robot control system also apply to the industrial robot control method 200 and the operations included therein, and will not be elaborated herein.

[0117] Next, refer to Figure 3 , which shows a schematic structural diagram of an electronic device 300 suitable for implementing some embodiments of the present invention. The electronic devices in some embodiments of the present invention may include, but are not limited to, mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Tablet Computers), PMPs (Portable Multimedia Players), in-vehicle terminals (such as in-vehicle navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc. Figure 3 The terminal device shown is only an example and should not impose any limitation on the functions and usage scope of the embodiments of the present invention.

[0118] As Figure 3 shown, the electronic device 300 may include a processing device (such as a central processing unit, a graphics processing unit, etc.) 301, which may perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 302 or the program loaded from the storage device 308 into the random access memory (RAM) 303. In the RAM 303, various programs and data required for the operation of the electronic device 300 are also stored. The processing device 301, the ROM 302, and the RAM 303 are connected to each other through a bus 304. The input / output (I / O) interface 305 is also connected to the bus 304.

[0119] Typically, the following devices can be connected to the I / O interface 305: input devices 306 including, for example, a touch screen, a touchpad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, etc.; output devices 307 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; storage devices 308 including, for example, magnetic tapes, hard disks, etc.; and a communication device 309. The communication device 309 can allow the electronic device 300 to communicate with other devices wirelessly or wiredly to exchange data. Although Figure 3 the electronic device 300 with various devices is shown, it should be understood that it is not required to implement or have all the shown devices. Instead, more or fewer devices can be implemented or had. Figure 3 Each block shown in [the figure] can represent one device or multiple devices as needed.

[0120] Furthermore, the storage medium of the embodiments of the present application stores program instructions capable of implementing all the above methods. Among them, the program instructions can be stored in the above storage medium in the form of a software product, including several instructions to cause a computer device (which can be a personal computer, a server, or a network device, etc.) or a processor to execute all or part of the steps of the methods described in the various embodiments of the present application. And the aforementioned storage medium includes: various media that can store program codes such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disc, or a terminal device such as a computer, a server, a mobile phone, a tablet, etc.

[0121] The above description is only some preferred embodiments of the present invention and an explanation of the applied technical principles. Those skilled in the art should understand that the scope of the invention involved in the embodiments of the present invention is not limited to the technical solutions formed by the specific combination of the above technical features, and should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the above inventive concept. For example, the technical solutions formed by mutually replacing the above features with the (but not limited to) technical features with similar functions disclosed in the embodiments of the present invention.

Claims

1. An industrial robot control system, characterized in that, It includes a trajectory planning module, a dynamic compensation module, a feedback control module, an execution module, a sensor module, and a data storage module; The trajectory planning module is used to generate the desired trajectory of the robot end effector according to the target pose; The dynamic compensation module is used to calculate the dynamic compensation torque based on the desired trajectory and the robot dynamics model; The feedback control module is used to generate the feedback control torque according to the deviation between the real-time pose data collected by the sensor module and the desired trajectory; The execution module is used to drive the robot joint movement according to the combined torque of the dynamic compensation torque and the feedback control torque; The sensor module is used to collect the robot joint angle, angular velocity, and torque data in real time; The data storage module is used to store the robot dynamics model parameters, control algorithm parameters, and historical motion data; The output ends of the dynamic compensation module and the feedback control module are both connected to the input end of the execution module.

2. The industrial robot control system according to claim 1, characterized in that, The trajectory planning module includes an inverse kinematics unit and an interpolation calculation unit. The inverse kinematics unit calculates the desired angles of each joint of the robot according to the target pose. The interpolation calculation unit generates a continuous and smooth joint angle trajectory through a quintic polynomial interpolation algorithm, and its trajectory generation formula is: q(t) = a0 + a1t + a2t 2 + a3t 3 + a4t 4 + a5t 5 where q(t) is the joint angle at time t, and a0 to a5 are interpolation coefficients determined by the initial angle, target angle, and boundary conditions.

3. The industrial robot control system according to claim 1, wherein The dynamic compensation module uses the Newton-Euler equation to calculate the robot dynamics compensation torque, and its calculation formula is: τ d = M(q)q + C(q, q)q + g(q) where τ d is the dynamic compensation torque, M(q) is the symmetric positive definite mass matrix, C(q, q) is the Coriolis force matrix, g(q) is the gravity vector, q is the joint angle vector, q is the joint angular velocity vector, and q is the joint angular acceleration vector.

4. The industrial robot control system according to claim 1, wherein, The feedback control module uses the PID control algorithm to generate the feedback control torque, and its calculation formula is: Among them, τ f is the feedback control torque, e is the deviation between the desired trajectory and the real-time pose, K p is the proportional coefficient, K i is the integral coefficient, K d is the differential coefficient.

5. The industrial robot control system according to claim 1, characterized in that The sensor module includes a joint encoder and a six-axis force sensor. The joint encoder is used to collect the joint angle and angular velocity, and the six-axis force sensor is used to detect the contact force of the end effector.

6. The industrial robot control system according to claim 1, characterized in that, The parameters stored in the data storage module include the robot link mass, center of mass position, inertia tensor, PID control coefficients, and dynamic compensation torque threshold.

7. The industrial robot control system according to claim 3, characterized in that, The dynamic compensation module also includes an adaptive compensation unit, which online updates the dynamics parameters through the Lyapunov stability theory, and the update formula is: Δθ = ΓY T s where Δθ is the parameter correction amount, Γ is a positive definite adaptive gain matrix, Y is a regression matrix, and s is a linear combination of the tracking errors.

8. The industrial robot control system according to claim 4, characterized in that, The PID coefficients of the feedback control module are dynamically adjusted through fuzzy logic rules, and its fuzzy rule table includes the membership functions of the error e and the error change rate and the output K p , K i , K d weighting factors.

9. The industrial robot control system according to claim 2, characterized in that, The interpolation calculation unit of the trajectory planning module uses the genetic algorithm to optimize the interpolation coefficients, and its fitness function is: F = w1∫(q) 2 dt + w2∫(q) 2 dt where w1 and w2 are weight coefficients, representing the optimization weights for the smoothness of acceleration and energy consumption respectively.

10. An industrial robot control method, characterized in that, It includes the following steps: S1. Generate the desired trajectory of the robot end effector through the trajectory planning module; S2. Calculate the dynamic compensation torque based on the desired trajectory and the dynamics model through the dynamic compensation module; S3. Generate the feedback control torque according to the deviation between the real-time pose and the desired trajectory through the feedback control module; S4. Combine the dynamic compensation torque and the feedback control torque and then drive the robot joint movement; S5. Collect the joint movement data in real time through the sensor module and feedback it to the dynamic compensation module and the feedback control module; S6. Update the dynamics model parameters and control algorithm parameters according to the historical motion data.

Images