A drive-control integrated control method and system

Abstract

This application discloses an integrated drive and control method and system, which relates to the field of industrial robot control technology. The disclosed integrated drive and control method and system are applied to a robotic arm. By predicting future current demand, assessing current saturation risk and constructing a rolling optimization problem, the joint position command is dynamically adjusted. This achieves multi-axis collaborative optimization while avoiding sudden current changes, thereby improving trajectory tracking accuracy, reducing mechanical vibration and realizing dynamic current optimization.

Description

Technical Field

[0001] This application relates to the field of industrial robot control technology, and in particular to an integrated drive and control method and system. Background Technology

[0002] Integrated drive and control technology, which tightly integrates the actuator and controller, is an important direction for the development of modern robotic arms. However, existing technologies face multiple technical bottlenecks in multi-axis collaborative motion control: First, traditional current loop limiting methods directly cut off commands when the joint motor reaches its maximum output current, leading to sudden changes in torque output, causing trajectory deviation and mechanical vibration; second, multi-axis collaborative control suffers from information transmission lag, making it difficult to adjust trajectory planning in time when joint current saturation is predicted; third, there is a lack of effective dynamic coordination mechanisms among the joint actuators, making it impossible to achieve global optimization of current demand allocation. Existing solutions are mostly static compensation or local adjustment, which cannot accurately predict changes in current demand over multiple control cycles and lacks collaborative optimization capabilities based on prediction results, leading to problems such as trajectory distortion, increased vibration, and performance waste in high-speed, high-precision robotic arms.

[0003] The above content is only used to help understand the technical solution of this application and does not represent an admission that the above content is prior art. Summary of the Invention

[0004] The main purpose of this application is to provide an integrated drive and control method and system, which aims to improve trajectory tracking accuracy, reduce mechanical vibration and achieve dynamic current optimization.

[0005] To achieve the above objectives, this application proposes an integrated drive and control method, which includes: The desired trajectory of the robot end effector is generated based on the externally input task instructions, and the desired position data and desired speed data of each joint for the current control cycle and multiple future control cycles are obtained based on the desired trajectory of the robot end effector; at the same time, the actual position data, actual speed data and actual current data output to each joint motor are obtained from the encoder of each joint. Based on the expected position data, the expected velocity data, the actual position data, the actual velocity data, and the actual current data, predictive data of the current demand of each joint in multiple future control cycles are obtained. Based on the actual current data and the pre-stored maximum allowable current data of each joint, the real-time current margin data of each joint is calculated; at the same time, combined with the current demand prediction data, the current saturation risk of each joint in the future multiple control cycles is assessed. Based on the expected position data, the current demand prediction data, and the current saturation risk, a rolling optimization problem data package is constructed with the goal of minimizing trajectory tracking error. The optimization variable is the optimized position command of each joint, and the constraints include the upper limit constraint of current and the dynamic model constraint. The rolling optimization problem data packet is solved by quadratic programming, and the optimized position command sequence for multiple future control cycles is output. The optimized position command data of the first cycle in the optimized position command sequence is extracted as the current control command and output to the joint actuator.

[0006] In one embodiment, the step of obtaining predicted current demand data for each joint over multiple future control cycles based on the expected position data, the expected velocity data, the actual position data, the actual velocity data, and the actual current data includes: The desired velocity data is subjected to time difference calculation to obtain the desired acceleration data for each joint; Based on a pre-defined joint dynamics model library, the rotational inertia data of the corresponding joint is called, and frictional force is estimated based on the actual velocity data to obtain frictional torque data. Based on the desired acceleration data, the moment of inertia data, and the frictional torque data, the required torque data for prediction is obtained through calculation and processing using the Newton-Euler equations. Based on a preset motor torque constant, the predicted torque data is converted into predicted current demand data through proportional conversion processing.

[0007] In one embodiment, the step of estimating friction force based on the actual speed data to obtain friction torque data includes: Based on actual speed data, speed direction identification data is obtained through sign function processing; Call the pre-stored joint coulomb friction coefficient and viscous friction coefficient; The friction torque data is obtained by multiplying the Coulomb coefficient of friction by the velocity direction indicator data, and then adding the product of the viscous friction coefficient and the actual velocity data.

[0008] In one embodiment, the step of assessing the current saturation risk of each joint over multiple future control cycles by combining the current demand prediction data includes: Based on the current demand forecast data, the compensated current demand data is obtained through time decay factor compensation processing; the time decay factor is set according to the current loop response bandwidth. The compensated current demand data is compared with the maximum allowable current data. When the compensated current demand data exceeds the maximum allowable current, the degree of current oversaturation is obtained by calculating the difference. Based on the degree of current oversaturation, a multi-level current saturation risk is output through threshold grading.

[0009] In one embodiment, the current saturation risk includes low-risk, medium-risk, and high-risk levels, and the step of outputting multi-level current saturation risk based on the degree of current oversaturation through threshold grading includes: When the current oversaturation level data is less than the preset first threshold, a low-risk level indicator data is output. When the current oversaturation level data is greater than the preset first threshold and less than the preset second threshold, output medium risk level identification data; When the current oversaturation level data is greater than the preset second threshold, a high-risk level indicator data and a corresponding joint indicator are output; wherein, the preset first threshold is less than the preset second threshold.

[0010] In one embodiment, the step of constructing a rolling optimization problem data package with the objective of minimizing trajectory tracking error based on the desired location data, the current demand prediction data, and the current saturation risk includes: For joints marked with a high risk level, a position change rate limit is added to the dynamic model constraints, which is achieved by setting a maximum allowable acceleration constraint; For joints identified as having a low risk level, their position tracking error weight is increased in the optimization objective.

[0011] In one embodiment, the step of performing quadratic programming on the rolling optimization problem data packet to output an optimized position command sequence for multiple future control cycles includes: The rolling optimization problem data package is converted into a standard quadratic programming matrix; The standard quadratic programming matrix is ​​processed by an effective set iterative algorithm to obtain a numerical solution for the optimized position command sequence; The numerical solution of the optimized position command sequence is subjected to command smoothing filtering to output an optimized position command sequence that satisfies the continuity of the robot's motion.

[0012] In one embodiment, the step of performing instruction smoothing filtering on the numerical solution of the optimized position command sequence to output an optimized position command sequence that satisfies the continuity of the robot's motion includes: The filter coefficient is set based on the rated acceleration parameter of the robotic arm; The optimized position command sequence is smoothed by the filtering coefficients described above. The slope of the optimized position command sequence after smoothing and filtering is checked to output an optimized position command sequence that satisfies the continuity of the robot's motion.

[0013] In one embodiment, when the robot is in a heavy-load direction-changing condition, before the step of constructing a rolling optimization problem data package with the objective of minimizing trajectory tracking error based on the desired position data, the current demand prediction data, and the current saturation risk, wherein the optimization variable is the optimized position command of each joint, and the constraints include current upper limit constraints and dynamic model constraints, the method further includes: Identify active directional joints based on historical motion data and obtain corresponding dynamic load inertia data; Based on the dynamic load inertia data, the maximum allowable current data of the corresponding joint is increased proportionally.

[0014] Furthermore, to achieve the above objectives, this application also proposes an integrated drive and control system, which includes: The data acquisition module is used to generate the desired trajectory of the robot end effector based on the externally input task instructions, and to acquire the desired position data and desired speed data of each joint in the current control cycle and multiple future control cycles based on the desired trajectory of the robot end effector; at the same time, it acquires the actual position data, actual speed data and actual current data output to each joint motor of each joint encoder. The current demand prediction module is used to obtain current demand prediction data for each joint in multiple future control cycles based on the expected position data, the expected speed data, the actual position data, the actual speed data, and the actual current data. The current saturation risk assessment module is used to calculate the real-time current margin data of each joint based on the actual current data and the pre-stored maximum allowable current data of each joint; at the same time, it combines the current demand prediction data to assess the current saturation risk of each joint in the future multiple control cycles. The joint position command optimization module is used to construct a rolling optimization problem data package with the goal of minimizing trajectory tracking error based on the expected position data, the current demand prediction data, and the current saturation risk. The optimization variable is the optimized position command of each joint, and the constraints include the current upper limit constraint and the dynamic model constraint. The output control module is used to perform secondary programming to solve the rolling optimization problem data packet, output the optimized position command sequence for multiple future control cycles, extract the optimized position command data of the first cycle in the optimized position command sequence as the current control command, and output it to the joint actuator.

[0015] The integrated drive and control method and system proposed in this application are applied to a robotic arm. By predicting future current demand, assessing current saturation risk and constructing a rolling optimization problem, the joint position command is dynamically adjusted. This achieves multi-axis collaborative optimization while avoiding sudden current changes, thereby improving trajectory tracking accuracy, reducing mechanical vibration and realizing dynamic current optimization. Attached Figure Description

[0016] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0017] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a flowchart illustrating an embodiment of the integrated drive and control method of this application. Figure 2 For this application Figure 1 A detailed flowchart of step S200; Figure 3 For this application Figure 2 A detailed flowchart of step S220; Figure 4 For this application Figure 1 Detailed flowchart of step S300; Figure 5 For this application Figure 4 A detailed flowchart of step S330; Figure 6 For this application Figure 1 Detailed flowchart of step S400; Figure 7 For this application Figure 1 Detailed flowchart of step S500; Figure 8 For this application Figure 7 A detailed flowchart of step S530; Figure 9 This is a flowchart illustrating another embodiment of the integrated drive and control method of this application. Figure 10 This is a schematic diagram of a structure provided for an embodiment of the integrated drive and control system of this application.

[0019] Explanation of icon numbers: 100. Integrated drive and control system; 110. Data acquisition module; 120. Current demand prediction module; 130. Current saturation risk assessment module; 140. Joint position command optimization module; 150. Output control module.

[0020] The purpose, features, and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0021] The technical solutions of this application will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of this application, and not all embodiments. The components of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0022] It should be understood that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this application, the terms "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0023] In existing technologies, integrated drive and control technology combines the actuator and controller, reducing size and communication latency. However, traditional solutions face rigid limitations due to current saturation during high-speed multi-axis coordinated motion, easily reaching physical limits when multiple joints require large current outputs simultaneously. Existing control architectures suffer from information transmission lag, with the trajectory adjustment layer unable to respond promptly to changes in current constraints, leading to insufficient instantaneous torque or trajectory deviation. While conservative strategies can avoid saturation, they sacrifice potential system performance. For example, when a robotic arm performs a high-speed arc-shaped motion, simultaneous acceleration of the shoulder and elbow joints may exceed their respective current limits. Traditional limiting methods directly cut off the current command, causing end-effector trajectory distortion and mechanical vibration.

[0024] To address the aforementioned issues, the inventors discovered that current saturation risk exhibits dynamic and multi-joint coupling characteristics. Traditional static limiting cannot adapt to load changes, and multi-axis coordinated control requires a forward-looking prediction mechanism. By analyzing the dynamic relationship between current demand and position commands, they proposed incorporating current predictions for multiple future control cycles into the optimization framework. Considering the balance between trajectory tracking and current constraints, a rolling optimization model needs to be established to adjust joint position commands while satisfying dynamic constraints. To address the delay issue, a quadratic programming method in the predictive time domain is employed to generate optimized command sequences in advance, ensuring the real-time performance and coordination of the control system.

[0025] Based on this, the embodiments of this application provide an integrated drive and control method applied to a robotic arm, referring to... Figure 1 The integrated drive and control method includes steps S100 to S500, wherein: Step S100: Generate the desired trajectory of the robot end effector according to the externally input task instructions, and obtain the desired position data and desired speed data of each joint in the current control cycle and multiple future control cycles based on the desired trajectory of the robot end effector; at the same time, obtain the actual position data, actual speed data and actual current data output to each joint motor of each joint encoder; Step S200: Based on the expected position data, the expected velocity data, the actual position data, the actual velocity data, and the actual current data, obtain the predicted current demand data for each joint in multiple future control cycles; Step S300: Based on the actual current data and the pre-stored maximum allowable current data of each joint, calculate the real-time current margin data of each joint; at the same time, combine the current demand prediction data to assess the current saturation risk of each joint in the future multiple control cycles. Step S400: Based on the expected position data, the current demand prediction data, and the current saturation risk, construct a rolling optimization problem data package with the goal of minimizing trajectory tracking error. The optimization variable is the optimized position command of each joint, and the constraints include the upper limit constraint of current and the dynamic model constraint. Step S500: Perform secondary programming to solve the rolling optimization problem data packet, output the optimized position command sequence for multiple future control cycles, extract the optimized position command data of the first cycle in the optimized position command sequence as the current control command, and output it to the joint actuator.

[0026] In this embodiment, the future multiple control cycles refer to the control phases of multiple consecutive discrete time intervals after the current moment. Specifically, this can be achieved using fixed-step or variable-step prediction in the time domain to establish a multi-step prediction model. Current demand prediction data is calculated using a dynamic model and friction estimation to determine the required current values ​​for each future cycle. Specifically, this can be obtained by combining the Newton-Euler equation with the motor torque constant conversion, used to predict the possibility of current saturation. Current saturation risk is a quantitative assessment of the probability that the joint will exceed the maximum allowable current in future control cycles. Specifically, this can be determined by comparing the predicted current with threshold levels, used to dynamically adjust and optimize constraints. The rolling optimization problem data package contains the mathematical expression of the objective function and constraints. Specifically, it can be constructed as a quadratic programming problem with inequality constraints, used to solve for the optimal position command online. Solving the quadratic programming problem converts the optimization problem into a standard matrix form. Specifically, it can use the effective set algorithm or interior-point method to calculate the numerical solution, ensuring that a feasible command sequence is obtained within a finite time.

[0027] In this embodiment, at the beginning of each control cycle, the integrated drive and control system 100 decomposes the desired trajectory of the end effector into a sequence of desired positions and velocities for each joint through inverse kinematics calculation. Simultaneously, it acquires the actual motion state fed back by the encoder and the current value output by the actuator. Based on the dynamic model and friction compensation algorithm, it predicts the current value required to maintain the desired acceleration over multiple future control cycles. It monitors the current margin of each joint in real time and assesses whether current over-limits will occur in the future based on the predicted values. When a potential saturation risk is detected, it adds acceleration constraints to the corresponding joint or adjusts the tracking error weights in the optimization model. By solving a constrained quadratic programming problem, it obtains an optimized position command sequence that balances trajectory accuracy and current limitations. Finally, it selects the optimization result of the first cycle as the current control output, achieving dynamic adjustment and early avoidance of current saturation.

[0028] In this embodiment, a correlation model between current demand and position commands is established through multi-step prediction. Saturation risks are proactively avoided at the trajectory generation layer, and constraints are dynamically adjusted based on real-time load to achieve adaptive optimization. Compared to local compensation methods, this application coordinates multi-joint movements through global optimization, maintaining a balanced current distribution across axes during high-speed motion. This effectively solves the current saturation problem during multi-axis coordinated motion, maintaining trajectory tracking accuracy while avoiding sudden torque changes. Through forward-looking prediction and rolling optimization, the position command sequence is adjusted in advance, eliminating the delay effects of traditional control architectures. The dynamic constraint adjustment mechanism ensures system safety while fully utilizing the current output capability of the driver, avoiding performance waste caused by conservative strategies. Multi-joint coordinated optimization ensures consistent movement across axes in complex motion scenarios, reducing the risk of mechanical vibration.

[0029] In one feasible implementation, refer to Figure 2 Step S200 includes steps S210 to S240, wherein: Step S210: Perform time difference calculation on the desired velocity data to obtain the desired acceleration data for each joint; Step S220: Based on the preset joint dynamics model library, call the rotational inertia data of the corresponding joint, and estimate the friction force based on the actual speed data to obtain the friction torque data; Step S230: Based on the desired acceleration data, the moment of inertia data, and the friction torque data, the required torque data for prediction is obtained by calculation and processing using the Newton-Euler equations; Step S240: Based on the preset motor torque constant, the predicted torque data is converted into current demand prediction data through proportional conversion processing.

[0030] In this embodiment, time difference calculation refers to calculating acceleration using numerical differentiation methods, which can be implemented using the central difference method or the backward difference method, to extract acceleration information from discrete velocity sampling data. Moment of inertia data refers to joint dynamic parameters stored in a preset model library, which can be obtained in real-time by querying the inertia value under the corresponding pose through joint angle feedback, reflecting the dynamic change of load inertia during the robot arm's movement. Friction estimation refers to nonlinear friction modeling based on velocity signals, which can be implemented using a Coulomb-viscous hybrid friction model. The velocity direction is extracted using a sign function and combined with a preset friction coefficient to calculate the real-time friction torque. The Newton-Euler equations are rigid body dynamics calculation equations, which can be implemented through forward dynamics calculations, correlating acceleration, inertia, and external torque to solve for the torque required to drive the joint. The motor torque constant is the conversion coefficient between motor torque and current, which can be obtained from the parameter manual of the robot arm's shutdown motor, used to convert theoretical torque requirements into actual current commands.

[0031] In this embodiment, during the acquisition of current demand prediction data, the desired velocity sequence is first subjected to discrete difference calculation, for example, the desired acceleration corresponding to each control cycle is calculated using the three-point difference formula. Then, the moment of inertia parameter corresponding to the current joint angle is extracted from a pre-established joint dynamics database. Simultaneously, based on the real-time acquired actual velocity signal, the motion direction is determined using a sign function, and the composite frictional torque, including static and viscous friction, is calculated. The desired acceleration, moment of inertia, and frictional torque are substituted into the Newton-Euler dynamics equations to solve for the predicted torque required to drive the joint motion. Finally, based on the torque-current proportionality coefficient in the motor characteristic parameters, the predicted torque is linearly mapped to the corresponding current demand value, forming a current demand prediction sequence for multiple future control cycles.

[0032] In this embodiment, by dynamically calling rotational inertia data and accurately calculating frictional torque, the load characteristics of the robotic arm under different postures and motion states can be accurately reflected. The friction model established by combining real-time speed data significantly improves the accuracy of torque prediction. Furthermore, this application achieves accurate prediction of the future current demand of the multi-joint system, providing reliable input data for subsequent current saturation risk assessment and trajectory optimization. The prediction method based on dynamic inertia calling and real-time friction compensation effectively overcomes the prediction deviation of traditional static models under varying load and direction conditions, making the predicted current demand values ​​closer to actual control requirements. By combining dynamic calculations with motor parameters, a complete mapping model from motion commands to current demand is established, laying the foundation for early identification and avoidance of current saturation risks.

[0033] In one feasible implementation, refer to Figure 3 Step S220 includes steps S221 to S223, wherein: Step S221: Based on the actual speed data, obtain the speed direction identification data through sign function processing; Step S222: Recall the pre-stored joint coulomb friction coefficient and viscous friction coefficient; Step S223: Multiply the Coulomb friction coefficient by the velocity direction indicator data, and then add the product of the viscous friction coefficient and the actual velocity data to obtain the friction torque data.

[0034] In this embodiment, the sign function processing refers to converting the positive and negative values ​​of velocity into direction identifiers. Specifically, a sign function can be used to map velocity data to +1 or -1 to distinguish the direction of friction. The Coulomb coefficient of friction is the static friction component independent of the velocity direction, which can be obtained through experimental calibration or theoretical calculation, and is used to characterize the frictional resistance when the joint is stationary or in low-speed motion. The viscous friction coefficient is the kinetic friction component proportional to the velocity magnitude, which can be obtained through dynamic testing or model fitting, and is used to characterize the frictional resistance when the joint is in high-speed motion. The velocity direction identifier data refers to the binary data reflecting the direction of joint movement, which can be generated by judging the positive and negative signs of the actual velocity using a sign function, and is used to determine the direction of the Coulomb friction force.

[0035] In this embodiment, when calculating the frictional torque, the sign of the actual velocity is first converted to a direction indicator using a sign function; for example, +1 is output when the velocity is positive and -1 is output when the velocity is negative. Then, the pre-stored Coulomb's coefficient of friction and viscous friction coefficient are called. The Coulomb's coefficient of friction is multiplied by the direction indicator to obtain the direction-dependent static friction component, while the viscous friction coefficient is multiplied by the actual velocity to obtain the velocity-proportional kinetic friction component. Finally, the two components are added together to obtain the total frictional torque data at the current velocity. This frictional torque data is used in subsequent torque prediction calculations to more accurately reflect the frictional effects on the joint during movement.

[0036] In this embodiment, by introducing direction identification data, the direction of Coulomb friction is dynamically correlated with the velocity direction. Combined with the velocity ratio characteristics of viscous friction, the friction characteristics of the joint under different motion states can be described more accurately. This allows for accurate differentiation between the contributions of static and dynamic friction, avoids torque prediction deviations caused by misjudgment of the friction direction, improves the accuracy of current demand prediction, reduces the risk of current saturation misjudgment caused by friction estimation errors, reduces trajectory tracking errors of the robotic arm during start-up, stop, or change-of-direction phases, and suppresses mechanical vibrations caused by sudden torque changes.

[0037] In one feasible implementation, refer to Figure 4 Step S300 includes steps S310 to S330, wherein: Step S310: Based on the current demand prediction data, the compensated current demand data is obtained through time decay factor compensation processing; the time decay factor is set according to the current loop response bandwidth. Step S320: Compare the compensated current demand data with the maximum allowable current data. When the compensated current demand data exceeds the maximum allowable current, calculate the degree of current oversaturation by the difference. Step S330: Based on the degree of current oversaturation, output multi-level current saturation risk through threshold grading.

[0038] In this embodiment, the time decay factor refers to the dynamic compensation coefficient used to correct the predicted current demand value. Specifically, it can be implemented using an exponential decay function, whose decay rate matches the closed-loop response speed of the current loop, compensating for the time difference between the control cycle and the current loop response delay. The compensated current demand data refers to the dynamically corrected predicted current demand value, specifically achieved by multiplying the original predicted data by the time decay factor, to more accurately reflect the actual current loop tracking capability. The current oversaturation level refers to the magnitude by which the predicted current demand exceeds the maximum allowable current, specifically calculated using the absolute value of the difference, used to quantify the severity of current overruns. Threshold grading refers to a classification method that divides risk levels based on the degree of oversaturation, specifically implemented using preset multi-level threshold values ​​to distinguish different levels of constraint violations.

[0039] In this embodiment, the current demand prediction data is first processed by a time decay factor compensation, for example, multiplying the predicted current by a decay factor in the range of 0.8 to 0.95, which is dynamically adjusted according to the current loop bandwidth. The compensated current demand is compared with the pre-stored maximum allowable current in real time. When an over-limit is detected, the absolute value of the excess is calculated as an oversaturation level indicator. This indicator is input into a three-level threshold classifier, for example, setting the first threshold to 5% of the maximum current and the second threshold to 15%, corresponding to low, medium, and high risk levels, respectively. When a high risk level is detected, the system will trigger a preventative control strategy adjustment.

[0040] In this embodiment, by introducing time decay factor compensation and multi-level threshold classification, the dynamic response characteristics of the current loop are considered, and a quantitative and graded assessment of future risks is achieved. This allows for the early identification of potential current saturation risk levels, providing graded early warning information for subsequent optimized control. Through dynamic compensation for prediction data errors and graded risk assessment, false alarms or missed alarms caused by single threshold judgments are avoided, and differentiated control strategies are matched for different risk levels, maximizing trajectory tracking performance while ensuring system safety.

[0041] In one feasible implementation, refer to Figure 5Step S330 includes steps S331 to S333, wherein: Step S331: When the current oversaturation level data is less than the preset first threshold, output the low risk level identification data; Step S332: When the current oversaturation level data is greater than a preset first threshold and less than a preset second threshold, output the medium risk level identification data; Step S333: When the current oversaturation level data is greater than the preset second threshold, output high-risk level identification data and corresponding joint identification; wherein, the preset first threshold is less than the preset second threshold.

[0042] In this embodiment, the current oversaturation level data refers to the difference between the predicted current demand and the maximum allowable current, which can be implemented through a real-time difference calculation module to quantify the degree to which the current exceeds the safe range. The preset first threshold is the critical value that triggers a low-risk level judgment, specifically set to 10%-20% of the maximum allowable current, used to identify minor over-limit conditions. The preset second threshold is the critical value that triggers a high-risk level judgment, specifically set to 30%-50% of the maximum allowable current, used to identify severe over-limit conditions. The joint identifier refers to the joint number information where high-risk saturation occurs, specifically implemented through an address encoder, used to accurately locate abnormal joints.

[0043] In this embodiment, during the high-speed movement of the robotic arm, when the predicted current demand for a certain joint in a future control cycle exceeds its maximum allowable current, the absolute difference between the two is first calculated as the oversaturation level. If the difference is below a first threshold, only a low-risk level is marked, and the original control strategy is maintained; if the difference is between the first and second thresholds, a medium-risk level is marked and a current compensation algorithm is initiated; if the difference exceeds the second threshold, a high-risk flag and joint number are immediately output, triggering an emergency constraint mechanism. For example, when the robotic arm executes a fast circular trajectory, if the second joint experiences a sudden load change causing the predicted current demand to reach 1.4 times the maximum allowable current, and the oversaturation level exceeds the second threshold, the system immediately generates a high-risk flag and locks the joint number, providing accurate fault location for subsequent optimization.

[0044] In this embodiment, a three-level risk classification is achieved by setting two threshold levels. Control performance is maintained at low risk, a compensation mechanism is activated at medium risk, and problematic joints are precisely located and subjected to enhanced constraints at high risk. This allows for accurate identification of different risk levels before current saturation occurs, enabling targeted hierarchical control strategies. In the low-risk stage, trajectory tracking accuracy is maintained; in the medium-risk stage, current demand is reduced through dynamic compensation; and in the high-risk stage, problematic joints are quickly located and motion constraints are applied. This ensures the safety of the robotic arm's movement while maximizing motion performance.

[0045] In one feasible implementation, refer to Figure 6 Step S400 includes steps S410 to S420, wherein: Step S410: For joints marked with a high risk level, add a position change rate limit to the dynamic model constraints, which is achieved by setting a maximum allowable acceleration constraint; Step S420: For joints identified as having a low risk level, increase their position tracking error weight in the optimization objective.

[0046] In this embodiment, referring to the above description, a high-risk level joint refers to a joint where the difference between the predicted current demand and the maximum allowable current exceeds a second threshold. This can be achieved by comparing the predicted current demand with a preset threshold in real time, and its function is to identify joints that may experience severe current saturation in advance. The position change rate limit refers to a hard constraint on joint acceleration, which can be limited by the upper limit of the acceleration term in the dynamic equation to prevent high-risk joints from exceeding their physical limits due to excessive acceleration demands. A low-risk level joint refers to a joint where the predicted current demand does not exceed a first threshold, which is achieved through a tiered threshold judgment. Its function is to identify joints that still have current margin. The position tracking error weight refers to the coefficient of the corresponding joint position deviation term in the optimization objective function, which can be achieved by adjusting the corresponding elements of the weight matrix in the quadratic programming problem, and is used to improve the trajectory tracking accuracy of low-risk joints.

[0047] In this embodiment, when constructing the rolling optimization problem, differentiated processing is implemented based on the current saturation risk level of each joint. For high-risk joints, an acceleration upper limit is embedded in the dynamic constraints, for example, limiting the acceleration term in the dynamic equation to no more than 80% of the rated value; for low-risk joints, the weight coefficient of the position error term in the objective function is increased to 1.5 times the standard value. During the optimization solution process, the acceleration limit of high-risk joints directly acts on the constraint matrix, while the error weight adjustment of low-risk joints is reflected in the Hessian matrix of the objective function, thereby realizing dynamic adjustment of the optimization strategy.

[0048] In this embodiment, precise local constraint adjustment is achieved through risk level classification. Based on real-time risk assessment and optimization parameter linkage, while ensuring that high-risk joints do not exceed limits, the current margin of low-risk joints is fully utilized to improve trajectory accuracy. This effectively balances current constraints and trajectory tracking performance in multi-joint cooperative motion scenarios. For high-risk joints, acceleration limitation avoids torque abrupt changes and vibrations caused by current saturation; for low-risk joints, increased error weighting can reduce accumulated trajectory deviation. For example, in circular trajectory motion, it can reduce inter-joint cooperative errors, thereby improving the smoothness of the end-effector trajectory.

[0049] In one feasible implementation, refer to Figure 7Step S500 includes steps S510 to S530, wherein: Step S510: Convert the rolling optimization problem data packet into a standard quadratic programming matrix; Step S520: Process the standard quadratic programming matrix using an effective set iterative algorithm to obtain a numerical solution for the optimized position command sequence; Step S530: Perform instruction smoothing filtering on the numerical solution of the optimized position instruction sequence to output an optimized position instruction sequence that satisfies the continuity of the robot's motion.

[0050] In this embodiment, the standard quadratic programming matrix refers to the standard mathematical expression of minimizing trajectory tracking error by converting the objective function and constraints into a quadratic programming problem. Specifically, this can be achieved by converting the objective function into a quadratic form matrix and the constraints into a linear inequality matrix. Its function is to transform complex multivariate optimization problems into a mathematically solvable form. The effective set iterative algorithm is a numerical computation method for solving constrained quadratic programming problems. Specifically, it is achieved by dynamically identifying the effective constraint set and iteratively updating candidate solutions. Its function is to accurately handle current upper limits and dynamic model constraints while ensuring computational efficiency. Command smoothing filtering refers to filtering the optimized position command to eliminate abrupt signals. Specifically, this can be achieved using a moving average filter or a low-pass filter. Its function is to ensure the smoothness of the robotic arm's movement by suppressing high-frequency noise.

[0051] In this embodiment, after the rolling optimization problem data packet is converted into a standard quadratic programming matrix, the effective set iterative algorithm dynamically identifies the current effective constraint set and updates candidate solutions, gradually approaching the optimal solution that satisfies all constraints. In each iteration, the algorithm prioritizes handling constraints corresponding to high-risk joints that may cause current saturation, thereby quickly adjusting the optimized position command. Subsequently, the numerical solution obtained is subjected to command smoothing filtering, and the filtering intensity is adjusted by the filtering coefficients so that the optimized position command can satisfy the current constraint while avoiding mechanical vibration caused by sudden changes in the command.

[0052] In this embodiment, by combining multi-cycle prediction with quadratic programming, joint optimization of current constraints and trajectory tracking is achieved, effectively reducing control delay. Furthermore, existing technologies often use gradient descent for optimization, which has low computational efficiency and struggles to handle abrupt constraints. The effective set iterative algorithm, however, can dynamically adjust the effective constraint set, significantly improving the solution speed. This allows for the rapid generation of optimized position commands that satisfy current constraints in multi-axis cooperative motion scenarios, avoiding torque abrupt changes and trajectory deviations caused by current saturation. Simultaneously, high-frequency noise in the optimized commands is eliminated through command smoothing filtering, ensuring continuous and smooth robotic arm movement, reducing the risk of mechanical vibration, and thus improving control accuracy and system reliability in highly dynamic tasks.

[0053] In one feasible implementation, refer to Figure 8 Step S530 includes steps S531 to S533, wherein: Step S531: Set the filter coefficient based on the rated acceleration parameter of the robot arm; Step S532: Perform smoothing filtering on the optimized position command sequence based on the filtering coefficients; Step S533: Perform slope verification on the optimized position command sequence after smoothing and filtering to output an optimized position command sequence that satisfies the continuity of the robot's motion.

[0054] In this embodiment, the filtering coefficient refers to a parameter used to adjust the smoothness of the instruction sequence. Specifically, it can be set using the rated acceleration parameter of the robotic arm joints, for example, by using the ratio of the rated acceleration to the square of the sampling period as the baseline value of the filtering coefficient. This parameter setting method ensures that the rate of change of the filtered instructions does not exceed the mechanical structure's tolerance. Smoothing filtering refers to performing a low-pass filtering operation on the discrete optimization instruction sequence. Specifically, it can be implemented using a moving average filter or a Butterworth filter to eliminate potential step jumps during the quadratic programming solution process. Slope verification verifies whether the difference between adjacent points in the filtered instruction sequence exceeds the allowable range. Specifically, this can be achieved by calculating the ratio of the positional difference between adjacent instruction points to the time interval and comparing it with a preset maximum allowable slope threshold, thereby avoiding the introduction of new abrupt changes due to filtering.

[0055] In this embodiment, after obtaining the discrete optimized position command sequence through quadratic programming, the filtering coefficient is first determined based on the rated acceleration parameters of each joint of the robotic arm. For example, for a joint with a rated acceleration of 5 m / s², the filtering coefficient can be set to 0.005 under a 1 ms control cycle. This coefficient is then used to perform a sliding window averaging process on the command sequence, making the acceleration changes between adjacent command points more gradual. After filtering, the processed command sequence needs to be verified point by point, calculating the deviation between the position command increment and the current actual position within each control cycle. When the increment at a certain point exceeds a preset threshold, the command value at that point is automatically adjusted to meet the maximum allowable jump range, thereby ensuring that the position commands received by the actuator both meet the optimization objective and are practically executable.

[0056] In this embodiment, by setting the filter coefficient of the rated acceleration constraint, the smoothness of the command is ensured to match the dynamic characteristics of the mechanical system. At the same time, the slope verification forms a dual guarantee mechanism, which effectively suppresses mechanical shock while maintaining the accuracy of trajectory tracking. This can eliminate high-frequency jitter components in the optimized command sequence and avoid the tremor phenomenon at the end of the robotic arm caused by abrupt changes in the discrete optimization solution. Meanwhile, the dynamic verification mechanism ensures that the smoothed command is smoothly connected with the real-time position state, which significantly improves the motion stability and control safety under heavy load and direction change conditions.

[0057] In one feasible implementation, when the robot arm is in a heavy-load direction-changing condition, before step S400, the method further includes steps S610 to S620, wherein: Step S610: Identify active steering joints based on historical motion data and obtain corresponding dynamic load inertia data; Step S620: Based on the dynamic load inertia data, increase the maximum allowable current data of the corresponding joint proportionally.

[0058] In this embodiment, the heavy-load reversal condition refers to the working state where the load mass at the end effector of the robotic arm exceeds the rated value and the direction of motion changes abruptly. Specifically, this can be triggered by a threshold for the rate of change of acceleration of the end effector. The active reversal joint refers to the joint that undertakes the main task of adjusting the direction of motion under this condition, and can be identified by correlation analysis between the change in joint angular velocity and the curvature of the end effector trajectory. Dynamic load inertia data refers to the change in equivalent rotational inertia caused by load inertia, which can be estimated by multiplying the load mass parameter by the joint acceleration. The proportional increase of the maximum allowable current data refers to linearly expanding the upper limit of the current according to the increase in load inertia, which can be dynamically adjusted using a predefined proportional coefficient.

[0059] In this embodiment, when the robotic arm performs a heavy-load directional change, the second derivative of the angular velocity of each joint is calculated by analyzing the joint motion trajectory over several control cycles. Joints with sudden changes in angular acceleration exceeding a set threshold are selected as active directional change joints. For the identified joints, the dynamic load inertia is calculated based on their current acceleration data and load mass parameters. For example, when the load mass is 5 kg and the instantaneous joint acceleration reaches 20 rad / s², the dynamic load inertia can be estimated as 100 kg·m². Based on the ratio of this inertia value to the reference inertia, for example, when the inertia increases by 1.5 times, the maximum allowable current of the corresponding joint is temporarily increased to 1.2 times the original value, thereby providing a current margin for sudden load changes.

[0060] In some specific implementations, historical motion data can be selected from the joint position sequence within the most recent three control cycles for differential calculation. Acceleration features are then extracted after fitting the joint motion curve using cubic spline interpolation. Dynamic load inertia calculation can combine end-effector force sensor data with the Jacobian matrix for inverse dynamic compensation. For example, when an inertial force of 10N is measured at the end-effector, the equivalent torque of the joint is calculated using the transpose of the Jacobian matrix. The adjustment ratio of the maximum allowable current can be set with a nonlinear mapping relationship; for example, an exponential growth strategy is activated when the load inertia exceeds twice the rated value.

[0061] In this embodiment, by dynamically identifying the joints with the most drastic load changes and specifically increasing their current limit, the motion control bandwidth is effectively expanded while ensuring system safety, avoiding energy waste caused by globally increasing the current limit. Thus, this application can accurately identify the joints with the most significant load changes and dynamically adjust their current capacity when the robotic arm performs heavy-load directional movements. This prevents amplified trajectory deviations due to current saturation and avoids performance losses caused by conservative current limits. This solution is particularly suitable for scenarios with sudden load changes in handling operations, such as when grasping objects of unknown mass, where the drive capability can be adaptively adjusted to maintain trajectory accuracy.

[0062] In this embodiment, the integrated drive and control method is applied to the robotic arm. By predicting future current demand, assessing current saturation risk, and constructing a rolling optimization problem, it dynamically adjusts joint position commands. This avoids sudden current changes while achieving multi-axis collaborative optimization, thereby improving trajectory tracking accuracy, reducing mechanical vibration, and realizing dynamic current optimization.

[0063] It should be noted that the above examples are only for understanding this application and do not constitute a limitation on the integrated drive and control method of this application. Any simple modifications based on this technical concept are within the protection scope of this application.

[0064] This application also provides an integrated drive and control system 100, see reference. Figure 10 The integrated drive and control system 100 includes: The data acquisition module 110 is used to generate the desired trajectory of the robot end effector according to the externally input task instructions, and to acquire the desired position data and desired speed data of each joint in the current control cycle and multiple future control cycles based on the desired trajectory of the robot end effector; at the same time, it acquires the actual position data, actual speed data and actual current data output to each joint motor of each joint encoder. The current demand prediction module 120 is connected to the data acquisition module 110 and is used to acquire the current demand prediction data of each joint in multiple future control cycles based on the expected position data, the expected speed data, the actual position data, the actual speed data and the actual current data. The current saturation risk assessment module 130 is connected to the current demand prediction module 120. It is used to calculate the real-time current margin data of each joint based on the actual current data and the pre-stored maximum allowable current data of each joint; at the same time, it combines the current demand prediction data to assess the current saturation risk of each joint in the future multiple control cycles. The joint position command optimization module 140 is connected to the current saturation risk assessment module 130. It is used to construct a rolling optimization problem data package with the goal of minimizing trajectory tracking error based on the expected position data, the current demand prediction data and the current saturation risk. The optimization variable is the optimized position command of each joint, and the constraints include the current upper limit constraint and the dynamic model constraint. The output control module 150 is connected to the joint position command optimization module 140. It is used to perform secondary planning and solution on the rolling optimization problem data packet, output the optimized position command sequence for multiple future control cycles, extract the optimized position command data of the first cycle in the optimized position command sequence as the current control command, and output it to the joint driver.

[0065] In this embodiment, the data acquisition module 110 is a hardware and software combination that collects the motion state and control commands of the robotic arm in real time through multi-source sensors. Specifically, it can be implemented using an FPGA chip with high-speed communication capabilities to synchronously acquire encoder feedback signals and planned trajectory data. The current demand prediction module 120 is a current estimation unit based on a dynamic model and friction compensation, specifically implemented using a recurrent neural network algorithm, to predict the current demand of each joint in multiple future control cycles. The current saturation risk assessment module 130 is a dynamic assessment unit combining real-time current margin and predicted demand, specifically implemented using a sliding window statistical method, to quantify the possibility of current exceeding limits at each joint. The joint position command optimization module 140 is a computational unit that constructs and solves multi-constraint optimization problems, specifically implemented using a quadratic programming solver based on Lagrange multipliers, to generate optimized commands that balance tracking accuracy and current constraints. The output control module 150 is an execution unit that converts the optimization results into drive signals, specifically implemented using a DAC circuit with filtering function to ensure the smoothness and real-time performance of command output.

[0066] In this embodiment, the data acquisition module 110 decomposes the desired trajectory of the end effector into the desired position and velocity sequence of each joint through inverse kinematics calculation, and simultaneously acquires the actual position, velocity, and actual current value output by the encoder via a high-speed bus. The current demand prediction module 120 calculates the joint acceleration demand based on the dynamic model, estimates the torque demand by combining it with the friction compensation model, and then converts it into a current prediction value through motor parameters. The current saturation risk assessment module 130 compares the predicted current with the preset maximum current in a time sequence, and identifies joints that may continuously exceed the limit through a sliding window statistical method. The joint position command optimization module 140 applies acceleration constraints to high-risk joints and increases the tracking weight for low-risk joints, constructing a multi-objective optimization model that includes current constraints and dynamic equations. The output control module 150 uses quadratic programming to obtain future multi-cycle optimized commands, which are then output to the driver after slope limiting filtering to ensure that command jumps are within the allowable range of the mechanical system.

[0067] The integrated drive and control system 100 provided in this application, employing the integrated drive and control method described in the above embodiments, can improve trajectory tracking accuracy, reduce mechanical vibration, and achieve dynamic current optimization. Compared with the prior art, the beneficial effects of the integrated drive and control system 100 provided in this application are the same as those of the integrated drive and control method provided in the above embodiments, and other technical features in the integrated drive and control system 100 are the same as those disclosed in the methods of the above embodiments, and will not be repeated here.

[0068] The above description is only a part of the embodiments of this application and does not limit the patent scope of this application. All equivalent structural transformations made under the technical concept of this application and using the contents of the specification and drawings of this application, or direct / indirect applications in other related technical fields, are included in the patent protection scope of this application.

Claims

1. A drive-control integrated control method, applied to a robotic arm, characterized in that, The aforementioned integrated drive and control method includes: The desired trajectory of the robot end effector is generated based on the externally input task instructions, and the desired position data and desired speed data of each joint for the current control cycle and multiple future control cycles are obtained based on the desired trajectory of the robot end effector; at the same time, the actual position data, actual speed data and actual current data output to each joint motor are obtained from the encoder of each joint. Based on the expected position data, the expected velocity data, the actual position data, the actual velocity data, and the actual current data, predictive data of the current demand of each joint in multiple future control cycles are obtained. Based on the actual current data and the pre-stored maximum allowable current data of each joint, the real-time current margin data of each joint is calculated; at the same time, combined with the current demand prediction data, the current saturation risk of each joint in the future multiple control cycles is assessed. Based on the expected position data, the current demand prediction data, and the current saturation risk, a rolling optimization problem data package is constructed with the goal of minimizing trajectory tracking error. The optimization variable is the optimized position command of each joint, and the constraints include the upper limit constraint of current and the dynamic model constraint. The rolling optimization problem data packet is solved by quadratic programming, and the optimized position command sequence for multiple future control cycles is output. The optimized position command data of the first cycle in the optimized position command sequence is extracted as the current control command and output to the joint actuator.

2. The integrated drive and control method as described in claim 1, characterized in that, The step of obtaining the predicted current demand data for each joint in multiple future control cycles based on the expected position data, the expected velocity data, the actual position data, the actual velocity data, and the actual current data includes: The desired velocity data is subjected to time difference calculation to obtain the desired acceleration data for each joint; Based on a pre-defined joint dynamics model library, the rotational inertia data of the corresponding joint is called, and frictional force is estimated based on the actual velocity data to obtain frictional torque data. Based on the desired acceleration data, the moment of inertia data, and the frictional torque data, the required torque data for prediction is obtained through calculation and processing using the Newton-Euler equations. Based on a preset motor torque constant, the predicted torque data is converted into predicted current demand data through proportional conversion processing.

3. The integrated drive and control method as described in claim 2, characterized in that, The step of estimating friction force based on the actual speed data to obtain friction torque data includes: Based on actual speed data, speed direction identification data is obtained through sign function processing; Call the pre-stored joint coulomb friction coefficient and viscous friction coefficient; The friction torque data is obtained by multiplying the Coulomb coefficient of friction by the velocity direction indicator data, and then adding the product of the viscous friction coefficient and the actual velocity data.

4. The integrated drive and control method as described in claim 1, characterized in that, The step of assessing the risk of current saturation for each joint in multiple future control cycles by combining the current demand forecast data includes: Based on the current demand forecast data, the compensated current demand data is obtained through time decay factor compensation processing; the time decay factor is set according to the current loop response bandwidth. The compensated current demand data is compared with the maximum allowable current data. When the compensated current demand data exceeds the maximum allowable current, the degree of current oversaturation is obtained by calculating the difference. Based on the degree of current oversaturation, a multi-level current saturation risk is output through threshold grading.

5. The integrated drive and control method as described in claim 4, characterized in that, The current saturation risk includes low-risk, medium-risk, and high-risk levels. The step of outputting multi-level current saturation risk based on the degree of current oversaturation through threshold grading includes: When the current oversaturation level data is less than the preset first threshold, a low-risk level indicator data is output. When the current oversaturation level data is greater than the preset first threshold and less than the preset second threshold, output medium risk level identification data; When the current oversaturation level data is greater than the preset second threshold, a high-risk level indicator data and a corresponding joint indicator are output; wherein, the preset first threshold is less than the preset second threshold.

6. The integrated drive and control method as described in claim 5, characterized in that, The steps for constructing a rolling optimization problem data package with the objective of minimizing trajectory tracking error, based on the desired location data, the current demand prediction data, and the current saturation risk, include: For joints marked with a high risk level, a position change rate limit is added to the dynamic model constraints, which is achieved by setting a maximum allowable acceleration constraint; For joints identified as having a low risk level, their position tracking error weight is increased in the optimization objective.

7. The integrated drive and control method as described in claim 1, characterized in that, The step of performing quadratic programming to solve the rolling optimization problem data packet and outputting the optimized position command sequence for multiple future control cycles includes: The rolling optimization problem data package is converted into a standard quadratic programming matrix; The standard quadratic programming matrix is ​​processed by an effective set iterative algorithm to obtain a numerical solution for the optimized position command sequence; The numerical solution of the optimized position command sequence is subjected to command smoothing filtering to output an optimized position command sequence that satisfies the continuity of the robot's motion.

8. The integrated drive and control method as described in claim 7, characterized in that, The step of performing instruction smoothing filtering on the numerical solution of the optimized position command sequence to output an optimized position command sequence that satisfies the continuity of the robot's motion includes: The filter coefficient is set based on the rated acceleration parameter of the robotic arm; The optimized position command sequence is smoothed by the filtering coefficients described above. The slope of the optimized position command sequence after smoothing and filtering is checked to output an optimized position command sequence that satisfies the continuity of the robot's motion.

9. The integrated drive and control method as described in claim 1, characterized in that, When the robotic arm is in a heavy-load direction-changing condition, before the step of constructing a rolling optimization problem data package with the objective of minimizing trajectory tracking error based on the expected position data, the current demand prediction data, and the current saturation risk, wherein the optimization variable is the optimized position command of each joint and the constraints include current upper limit constraints and dynamic model constraints, the method further includes: Identify active directional joints based on historical motion data and obtain corresponding dynamic load inertia data; Based on the dynamic load inertia data, the maximum allowable current data of the corresponding joint is increased proportionally.

10. An integrated drive and control system, characterized in that, The integrated drive and control system includes: The data acquisition module is used to generate the desired trajectory of the robot end effector based on the externally input task instructions, and to acquire the desired position data and desired speed data of each joint in the current control cycle and multiple future control cycles based on the desired trajectory of the robot end effector; at the same time, it acquires the actual position data, actual speed data and actual current data output to each joint motor of each joint encoder. The current demand prediction module is used to obtain current demand prediction data for each joint in multiple future control cycles based on the expected position data, the expected speed data, the actual position data, the actual speed data, and the actual current data. The current saturation risk assessment module is used to calculate the real-time current margin data of each joint based on the actual current data and the pre-stored maximum allowable current data of each joint; at the same time, it combines the current demand prediction data to assess the current saturation risk of each joint in the future multiple control cycles. The joint position command optimization module is used to construct a rolling optimization problem data package with the goal of minimizing trajectory tracking error based on the expected position data, the current demand prediction data, and the current saturation risk. The optimization variable is the optimized position command of each joint, and the constraints include the current upper limit constraint and the dynamic model constraint. The output control module is used to perform secondary programming to solve the rolling optimization problem data packet, output the optimized position command sequence for multiple future control cycles, extract the optimized position command data of the first cycle in the optimized position command sequence as the current control command, and output it to the joint actuator.

Images