A control system for a six-joint robot

Abstract

The application relates to the technical field of robots, and particularly discloses a control system of a six-joint robot, which comprises a robot body, a detection unit, a control cabinet and a servo driving unit. The detection unit collects the positions and speeds of all joints, a controller in the control cabinet constructs a layered error according to a preset reference trajectory, generates virtual control and benchmark control input by adopting fixed-time convergence backstepping, estimates equivalent disturbance based on the residual signal of a nominal dynamic model and actual response, generates feedforward compensation and benchmark control input, and superimposes the benchmark control input to form a final control input vector to drive all joints. The system realizes offset in the initial stage of disturbance, reduces high-frequency oscillation and chattering of the control input near the steady state, improves trajectory tracking accuracy, and reduces servo heating, wear and noise.

Description

Technical Field

[0001] This invention relates to the field of robotics, and more specifically, to a control system for a six-joint robot. Background Technology

[0002] Six-joint robots typically require their end effectors to follow a preset reference trajectory in tasks such as welding, handling, and assembly. Existing programmable robotic arms generally employ a controller within a control cabinet to perform closed-loop trajectory tracking control of each joint, outputting joint torque or equivalent control quantities through servo drive units. However, in actual operation, robot dynamics are subject to uncertainties such as modeling errors, friction variations, load changes, and external disturbances. These uncertainties can easily cause high-frequency fluctuations in the control input near the steady state, manifesting as high-frequency oscillations or chattering. This can lead to problems such as servo overheating, mechanical wear, increased noise, and decreased tracking accuracy.

[0003] To improve convergence speed and tracking performance, academia and industry have proposed robot control methods with fixed-time / finite-time convergence. For example, in 2022, Meng Xianyang published a paper in the journal *Control Theory & Applications* on trajectory tracking control of an uncertain robotic arm based on contraction backstepping. This paper uses a recursive design within a "contraction backstepping" framework, making the error convergence time independent of initial conditions. Public research shows that using fixed-time inversion control and performance constraint design can enable the robotic arm error to converge according to preset performance and achieve smooth and bounded control torque, thereby avoiding the chattering problem inherent in traditional sliding mode control.

[0004] On the other hand, to combat uncertainties and external disturbances, disturbance observers and their improved forms are used in robot arm control. By estimating and compensating for unknown disturbance torques, they can be used for friction compensation, independent joint control, and robust control. Among existing publicly available solutions, Yang Peng published a sliding mode control method for the end effector of a robotic arm based on a nonlinear disturbance observer in the Journal of Zhengzhou University in 2019. Figure 2 As shown, a nonlinear disturbance observer is used to estimate online and the observed values ​​are superimposed on the sliding mode controller as feedforward compensation to reduce control chattering and achieve finite-time convergence of tracking error.

[0005] Despite existing research and patented solutions such as fixed-time convergence control and disturbance observation compensation, a type of engineering pain point still exists when implementing a six-joint robot control cabinet: when disturbances such as load changes occur in the early stages of their effect, if the main reliance is on error feedback to "catch up with the disturbance", high gain and rapid correction are often required, which in turn induces high-frequency control input fluctuations near the steady state. Summary of the Invention

[0006] To overcome the aforementioned deficiencies of the prior art, the present invention provides a control system for a six-joint robot. This system constructs hierarchical errors and generates virtual control quantities through a controller, and implements fixed-time convergence inverse control within the control cabinet. Combined with online disturbance estimation and feedforward compensation, the system cancels out disturbances in the early stages of their effect, reducing high-frequency oscillations and chattering phenomena of the control input near the steady state, thereby solving the problems mentioned in the background art.

[0007] To achieve the above objectives, the present invention provides the following technical solution:

[0008] A control system for a six-joint robot includes:

[0009] The robot body performs trajectory tracking motion of its six joints under the drive of the servo drive unit.

[0010] The detection unit collects position and velocity signals from the six joints.

[0011] The control cabinet generates a reference control input vector based on fixed-time inversion control, and combines it with online disturbance estimation to generate feedforward compensation, forming the final control input vector output to the servo drive unit.

[0012] The servo drive unit drives the joint actuators of the six-joint robot body to output joint torque and control trajectory tracking by means of joint channels.

[0013] As a further aspect of the present invention, the robot body performs trajectory tracking motion of six joints under the drive of the servo drive unit, including the following specific content: the robot body, as the controlled object and the carrier of the execution mechanism, contains six joints, which are used to enable the end effector to complete the trajectory tracking action along a preset reference trajectory.

[0014] As a further aspect of the present invention, the detection unit collects position and velocity signals of six joints, including the following specific content: the detection unit is used to collect feedback on the operating status of the six joints, and includes position sensors and velocity sensors corresponding to each of the six joints, and outputs position and velocity signals of the six joints.

[0015] As a further aspect of the present invention, a control cabinet generates a reference control input vector based on fixed-time inversion control, and generates a feedforward compensation quantity by combining online disturbance estimation to form a final control input vector output to the servo drive unit. The specific contents include: the control cabinet has a built-in controller. In each control cycle, the controller reads the desired position and desired velocity from a preset reference trajectory, and then calculates the position tracking error and velocity tracking error for each joint, forming a hierarchical error to characterize the current tracking deviation, and recursively generates a virtual control quantity sequence; using the virtual control quantity sequence as the control target for each joint, the virtual control quantity sequence is recursively corrected using the velocity error, and a six-dimensional reference control input vector is output; in each control cycle, the controller calculates the deviation between the actual joint position and the predicted joint position, and between the actual joint velocity and the predicted joint velocity for each joint, and uses the deviation as a residual signal; the residual signal is input to an observer, which maintains a set of time-updated disturbance estimates for each joint channel, outputs an equivalent disturbance estimate, converts it into a feedforward compensation quantity, and algebraically superimposes the feedforward compensation quantity with the reference control input vector to form a final control input vector output to the servo drive unit.

[0016] The controller outputs the final control input vector to the servo drive unit, which uses it as a control command for the six joints and applies it to the robot body, so that each joint can track the preset reference trajectory under the drive of the final control input vector.

[0017] The final control input vector is superimposed with a feedforward compensation amount on top of the reference control input vector. When the load change disturbance begins to take effect and causes the actual joint response to deviate from the model response, the compensation term generated by the controller participates in the output in the early stage of the disturbance to offset the additional effects caused by the disturbance.

[0018] The actual joint response refers to the actual joint position and the actual joint velocity. The actual joint position is obtained by measuring a position sensor, and the actual joint velocity is obtained by measuring a velocity sensor.

[0019] The model response is generated by the controller calling the internal model, which includes: the internal model is the nominal dynamic model of the six-joint robot body, which outputs the predicted joint position and predicted joint velocity under the condition that it is consistent with the control input output to the servo drive unit in the previous control cycle.

[0020] The preset reference trajectory is pre-stored in the storage medium of the controller, and the preset reference trajectory contains at least six joints corresponding to the desired position sequence and desired velocity sequence at each control moment.

[0021] As a further aspect of the present invention, the servo drive unit drives the joint actuators of the six-joint robot body to output joint torques according to the joint channels, controlling trajectory tracking. This includes the following specific details: The servo drive unit is used to translate the final control input generated by the controller into actual drive for each joint. After receiving the final control input vector, the servo drive unit outputs drive actions according to the servo drive channels corresponding to each of the six joints, causing each joint actuator to generate joint torque or an equivalent control quantity with the same dimensions, thereby driving the six-joint robot to move along a preset reference trajectory.

[0022] As a further aspect of the present invention, a control system for a six-joint robot can be used as an industrial robot on various production lines; as an example of an application scenario, the control system for the six-joint robot described in the present invention can be applied to industrial production scenarios such as welding, handling, and assembly.

[0023] The technical effects and advantages of the control system for a six-joint robot of this invention are as follows: This invention addresses the problem that six-joint robots are prone to high-frequency fluctuations in control input near steady state under uncertainties such as modeling errors, friction variations, load changes, and external disturbances, leading to servo heating, mechanical wear, increased noise, and decreased tracking accuracy. The controller in the control cabinet constructs layered errors such as position error layers and velocity error layers and recursively generates a virtual control quantity sequence. Fixed-time convergence inversion control is achieved within the control cabinet to form a six-dimensional reference control input. Simultaneously, the nominal dynamic model is called to obtain the model response, and the residual signal is calculated by comparing it with the actual joint response and input into the observer to obtain... The equivalent disturbance estimates for each joint are then used to generate feedforward compensation values ​​corresponding to the joint channels based on a preset mapping relationship. These compensation values ​​are then aligned with the reference control input within the same control cycle and superimposed on the joint channels to form the final control input vector, which is output to the servo drive unit. This allows the compensation terms to participate in the output at the initial stage of the disturbance to offset the additional effects of the disturbance. As a result, deviation suppression no longer relies primarily on increasing the feedback gain for rapid correction, reducing the dependence on high-gain, fast-changing feedback regulation. This makes the correction action near steady state more gentle, the control input change smoother, and reduces the accumulation of high-frequency components. Consequently, it reduces the high-frequency oscillation and chattering frequency of the control input near steady state and improves the trajectory tracking effect. Attached Figure Description

[0024] Figure 1 This is a schematic diagram of the control system of a six-joint robot according to the present invention.

[0025] Figure 2 This is a schematic diagram of the control input for a prior art device with an interference observer.

[0026] Figure 3 This is a schematic diagram of the six-joint robot body structure of the present invention.

[0027] Figure 4This is a schematic diagram showing the range of motion of the robot body of the present invention.

[0028] Figure 5 This is a schematic diagram of the control cabinet of the present invention.

[0029] Figure 6 This is a line graph showing the suppression of high-frequency oscillations near the steady state control input in this invention. Detailed Implementation

[0030] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0031] Example 1

[0032] like Figure 1 As shown, a control system for a six-joint robot includes:

[0033] The robot body performs trajectory tracking motion of its six joints under the drive of the servo drive unit.

[0034] The detection unit collects position and velocity signals from the six joints.

[0035] The control cabinet generates a reference control input vector based on fixed-time inversion control, and combines it with online disturbance estimation to generate feedforward compensation, forming the final control input vector output to the servo drive unit.

[0036] The servo drive unit drives the joint actuators of the six-joint robot body to output joint torque and control trajectory tracking by means of joint channels.

[0037] Furthermore, the robot body, driven by the servo drive unit, performs trajectory tracking motion of its six joints, including: In this embodiment, the robot body serves as both the controlled object and the carrier of the execution mechanism, such as... Figure 3 As shown, it contains six joints, which are used to enable the end effector to complete trajectory tracking action along a preset reference trajectory. Figure 4 The range of motion of the robot body is shown. The servo drive unit receives the final control input from the controller in the control cabinet and drives the joint actuators to output joint torque, thereby causing the robot body to produce joint movements and form actual joint position and speed responses.

[0038] Furthermore, the detection unit collects position and velocity signals of the six joints, including: the detection unit is used to collect feedback on the operating status of the six joints, and includes position sensors and velocity sensors corresponding to each of the six joints, and outputs position and velocity signals of the six joints.

[0039] The position sensor is installed at the angular position measurement point of each joint, based on the coaxial or rigidly coupled part of the joint's rotation axis, thereby directly reflecting the actual angular position of the joint; the velocity sensor is installed at the angular velocity measurement point coaxial with the position measurement point, to obtain the actual angular velocity of the joint under the combined action of control input and disturbance.

[0040] Furthermore, the control cabinet generates a reference control input vector based on fixed-time inversion control, and generates a feedforward compensation amount by combining online disturbance estimation to form a final control input vector output to the servo drive unit. The control cabinet includes a controller installed inside the control cabinet, which is signal-connected to the detection unit and the servo drive unit respectively. Figure 5 The structure of the control cabinet is shown.

[0041] The controller, as a computing and control unit located inside the control cabinet, stores a preset reference trajectory in its storage medium in advance (the preset reference trajectory refers to a target motion trajectory determined in advance to enable the six-joint robot to complete a certain trajectory tracking action). The preset reference trajectory contains at least the expected position sequence and expected velocity sequence of the six joints at each control moment, as the target input for trajectory tracking control. Within each control cycle, the controller first reads the desired position and desired velocity corresponding to the current moment from the preset reference trajectory. Then, the controller first extracts the desired position and desired velocity of each joint at the current moment from the preset reference trajectory (forming a set of desired position and desired velocity quantities), and then reads the actual joint position and actual joint velocity of each joint obtained by the detection unit synchronously sampling (forming a set of actual joint position and actual joint velocity quantities). Subsequently, the controller calculates the "difference between the desired position and the actual joint position" for each joint to obtain the position tracking error of that joint, and calculates the "difference between the desired velocity and the actual joint velocity" to obtain the velocity tracking error of that joint. The above position tracking errors and velocity tracking errors of the six joints are combined in the order of the joint channels to form a layered error used to characterize the current tracking deviation. The layered error includes a position error layer and a velocity error layer.

[0042] The controller divides the system into two layers of virtual controlled sub-links from the outside in, based on error levels. Taking the position error layer as the starting point for the first layer of recursion, it first constructs a target error evolution form that meets the requirements of a fixed-time convergence structure for the position error layer, and generates the first layer of virtual control quantity accordingly. Under the action of this virtual control quantity, the position error layer evolves according to the preset fixed-time convergence structure. Subsequently, the deviation between the actual joint speed and the first layer of virtual control quantity is defined as the speed error layer, and this speed error layer is used as the next recursive object. In the second layer of recursion, the speed error layer and the error coupling term of the previous layer are incorporated into a unified recursive construction process. Through the layer-by-layer updating and consistency processing of the virtual control quantity, the second layer error also meets the requirements of a fixed-time convergence structure evolution. In this recursive process, the virtual control quantity generated at each layer serves as a "reference target" for the definition of the error and the construction of the control law at the next layer, realizing a layer-by-layer "inversion elimination-recursive closed loop" from the position error layer to the speed error layer, forming a sequence of virtual control quantities used to generate the reference control input. The controller uses a virtual control quantity sequence as the control target for each joint, and uses speed error to recursively correct the virtual control quantity sequence, outputting a six-dimensional reference control input vector.

[0043] In each control cycle, the controller acquires the actual joint response output by the detection unit, which refers to the actual joint position and actual joint velocity. Simultaneously, the controller calls its internal model to generate a model response: the internal model is a nominal dynamic model of the six-joint robot body, which outputs predicted joint position and predicted joint velocity (the predicted joint position and predicted joint velocity are model responses) under conditions consistent with the control input actually applied to the servo drive unit in the previous control cycle. The controller calculates the deviation between "actual joint position - predicted joint position" and "actual joint velocity - predicted joint velocity" for each joint, and uses these deviations as residual signals.

[0044] The controller inputs the residual signal into the observer, which maintains a set of disturbance estimates that are updated over time for each joint channel. The observer, based on a preset observation structure and gain setting, uses the current residual signal to correct the disturbance estimate from the previous cycle, causing the disturbance estimate to converge in a direction that "can explain the additional effects of the deviation," and outputs the equivalent disturbance estimate for the current control cycle. If a disturbance observer is used, the equivalent disturbance moment estimate is recursively updated and output based on the unmodeled effects reflected in the residuals. If an extended state observer is used, uncertainties and external disturbances are treated as lumped terms and corrected in real time as the residuals change during observation updates. If a learning observer is used, the lumped disturbance effect is learned online under the drive of the residual signal, and the equivalent disturbance estimate corresponding to each joint is output to characterize the equivalent moment effect of the disturbance on the joint channel.

[0045] The controller obtains equivalent disturbance estimates for each of the six joints and generates feedforward compensation amounts for the corresponding joints based on a preset mapping relationship. The mapping relationship includes at least a proportional mapping or a gain matrix mapping, that is, a compensation amount with the same dimensions as the joint torque is obtained by applying a preset proportional coefficient or a preset gain matrix to the equivalent disturbance estimate.

[0046] The controller uses a reference control input vector as the basic control quantity and aligns the reference control input vector with the feedforward compensation quantity within the same control cycle, ensuring a one-to-one correspondence between the two in terms of joint number and time point. Subsequently, the controller performs algebraic superposition operations using the six joints as channels, that is, adding the feedforward compensation quantity of each joint to the reference control input component of that joint to obtain the superposition result for that joint; the superposition results of the six joints are then combined into a final control input vector according to a predetermined channel order. Through the above channel-based superposition and vectorized combination, the final control input vector matches the servo drive channels one-to-one in form and dimension, thus allowing it to be directly output to the servo drive unit as control commands for the six-joint actuator.

[0047] The controller outputs the final control input vector to the servo drive unit, which uses it as control commands for the six joints and applies them to the six-joint robot body. This allows each joint to track a preset reference trajectory under the drive of the final control input vector. Because the final control input vector includes a feedforward compensation for equivalent disturbances superimposed on the reference control input vector, when disturbances such as load changes begin to occur and cause the actual joint response to deviate from the model response, the compensation term generated by the controller can participate in the output at the initial stage of the disturbance, offsetting the additional effects caused by the disturbance. This means that deviation suppression on the joint channels no longer primarily relies on the error feedback channel increasing feedback gain and performing rapid correction, thereby reducing the dependence on high-gain, rapidly changing feedback adjustment. In this case, the controller does not need to continuously output control quantities with rapidly changing amplitudes and frequent adjustments to quickly suppress deviations, thus reducing the rapid fluctuation of the control input near the steady state between adjacent control cycles. In other words, after the disturbance effect is preferentially absorbed by the compensation term, the correction action required by the final control input vector near the steady state becomes more gentle, the change of the final control input vector is smoother, and the accumulation of high-frequency components caused by high-gain rapid correction is reduced, thereby reducing high-frequency oscillations and chattering phenomena of the control input near the steady state. Figure 6As shown, when a six-joint robot is subjected to external disturbances, relying solely on error feedback for rapid correction often results in more frequent and denser high-frequency fluctuations in the control input near the steady state between adjacent control cycles. However, this invention superimposes a feedforward compensation amount based on residuals and observers on top of the reference control input. This allows the compensation term to participate in the output in the early stages of the disturbance and preferentially absorb the disturbance's influence, thereby reducing the dependence on high-gain, fast-changing feedback regulation. This makes the correction action near the steady state gentler, the control input change smoother, and the accumulation of high-frequency components reduced, ultimately resulting in a decrease in high-frequency oscillations and chattering near the steady state.

[0048] Furthermore, the servo drive unit, by driving the joint actuators of the six-joint robot body to output joint torques according to the joint channels, controls trajectory tracking, including: the servo drive unit is used to implement the final control input generated by the controller into actual drive for each joint. After receiving the final control input vector, the servo drive unit outputs drive actions according to the servo drive channels corresponding to each of the six joints, causing each joint actuator to generate joint torque or an equivalent control quantity with the same dimensions, thereby driving the six-joint robot to move along a preset reference trajectory.

[0049] Example 2

[0050] This embodiment, without altering the structure of the six-joint robot control system described in Embodiment 1, presents a typical operating condition: when the six-joint robot performs handling or assembly tasks, the end effector experiences load changes during trajectory tracking, causing the actual response on the joint channel to deviate from the predicted response of the nominal dynamic model.

[0051] The system still includes the robot body, detection unit, control cabinet and servo drive unit, and the connection relationship and signal interaction method of each module are the same as in Example 1.

[0052] The controller inside the control cabinet maintains a fixed control cycle. At the beginning of each control cycle, the controller reads the desired positions and velocities of the six joints corresponding to the current control moment from the preset reference trajectory, and simultaneously reads the actual joint positions and velocities of the six joints collected by the detection unit. The controller calculates the position tracking error and velocity tracking error for each joint separately, and combines them according to the joint channel sequence to form a layered error, wherein the layered error includes at least a position error layer and a velocity error layer.

[0053] Subsequently, the controller performs recursion within the fixed-time inversion control framework: starting from the position error layer, it constructs a target error evolution form that satisfies the fixed-time convergence structure requirements, generates the first layer of virtual control quantity, and uses the deviation between the actual joint speed and the first layer of virtual control quantity as the speed error layer, continuing to perform the next layer of recursive correction, so that the errors of each layer satisfy the fixed-time convergence structure evolution requirements, thereby forming a sequence of virtual control quantities used to generate the reference control input.

[0054] The controller uses a virtual control sequence as the control target for each joint, and recursively corrects the virtual control sequence using velocity error, outputting a six-dimensional reference control input vector. Simultaneously, the controller calls its internal model (nominal dynamics model) to output predicted joint positions and predicted joint velocities as model responses, under the condition that the control input actually applied to the servo drive unit in the previous control cycle is consistent with the model's response. The controller uses the actual joint positions and velocities obtained from the detection unit as the actual joint responses, and calculates the deviations between "actual joint position - predicted joint position" and "actual joint velocity - predicted joint velocity," forming residual signals. After the residual signals are input to the observer, the observer maintains a set of time-updated disturbance estimates for each joint channel, and corrects the disturbance estimates of the previous cycle based on the current residual signals, outputting the equivalent disturbance estimate for the current cycle.

[0055] Subsequently, the controller converts the equivalent disturbance estimate into a feedforward compensation value that corresponds one-to-one with the joint channel according to the preset mapping relationship (at least including proportional mapping or gain matrix mapping), and aligns the feedforward compensation value with the reference control input vector in the same control cycle to ensure that the joint number corresponds one-to-one with the time point.

[0056] Next, the controller performs algebraic superposition of the two along the joint channels: for each joint, the feedforward compensation amount of that joint is added to the reference control input of that joint to form the final control amount of that joint. The superposition results of the six joints are combined into a final control input vector and output to the servo drive unit. The servo drive unit drives the actuators of each joint along the joint channels to output joint torque or equivalent control amount, thereby driving the robot body to move along a preset reference trajectory.

[0057] Since the feedforward compensation comes from the online estimation of disturbance driven by the residual signal, it can participate in the output at the beginning of the load change disturbance and the actual response deviates from the model response. It can offset the additional effects caused by the disturbance, so that deviation suppression no longer mainly depends on increasing the feedback gain and fast correction. This reduces the rapid fluctuation of the control input near the steady state between adjacent control cycles, reduces the tendency of high-frequency oscillation or chattering, and improves the trajectory tracking effect.

[0058] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0059] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A control system for a six-joint robot, characterized in that, include: The robot body performs trajectory tracking motion of its six joints under the drive of the servo drive unit; The detection unit collects position and velocity signals from six joints; The control cabinet generates a reference control input vector based on fixed-time inversion control, combines online disturbance estimation to generate feedforward compensation, and forms the final control input vector output to the servo drive unit. The servo drive unit drives the joint actuators of the six-joint robot body to output joint torque and control trajectory tracking according to the joint channel; The control cabinet has a built-in controller. In each control cycle, the controller reads the desired position and desired speed from the preset reference trajectory, and then calculates the position tracking error and speed tracking error for each joint to form a hierarchical error to characterize the current tracking deviation, and recursively generates a sequence of virtual control quantities. Using the virtual control quantity sequence as the control target for each joint, the speed error is used to recursively correct the virtual control quantity sequence, and a six-dimensional reference control input vector is output. In each control cycle, the controller calculates the deviation between the actual joint position and the predicted joint position, and between the actual joint speed and the predicted joint speed for each joint, and uses the deviation as a residual signal. The residual signal is input to the observer, which maintains a set of disturbance estimates that are updated over time for each joint channel, outputs an equivalent disturbance estimate, converts it into a feedforward compensation quantity, and algebraically superimposes the feedforward compensation quantity with the reference control input vector to form the final control input vector, which is output to the servo drive unit.

2. The control system for the six-joint robot according to claim 1, characterized in that... The controller outputs the final control input vector to the servo drive unit, which uses it as a control command for the six joints and applies it to the robot body, so that each joint can track the preset reference trajectory under the drive of the final control input vector.

3. The control system for the six-joint robot according to claim 2, characterized in that... The final control input vector is superimposed with a feedforward compensation amount in addition to the reference control input vector. When the load change disturbance begins to take effect and causes the actual joint response to deviate from the model response, the feedforward compensation amount is generated by the equivalent disturbance estimate through a preset mapping relationship, and is superimposed and output according to the joint channel after being time-aligned with the reference control input vector.

4. The control system for the six-joint robot according to claim 3, characterized in that, The actual joint response refers to the actual joint position and the actual joint velocity. The actual joint position is obtained by measuring a position sensor, and the actual joint velocity is obtained by measuring a velocity sensor.

5. The control system for the six-joint robot according to claim 3, characterized in that, The model response is generated by the controller calling the internal model, which includes: the internal model is the nominal dynamic model of the six-joint robot body, which outputs the predicted joint position and predicted joint velocity under the condition that it is consistent with the control input output to the servo drive unit in the previous control cycle.

6. The control system for the six-joint robot according to claim 2, characterized in that, The preset reference trajectory is pre-stored in the storage medium of the controller, and the preset reference trajectory contains at least six joints corresponding to the desired position sequence and desired velocity sequence at each control moment.

7. The control system for the six-joint robot according to claim 1, characterized in that, The detection unit includes position sensors and velocity sensors corresponding to each of the six joints, and outputs position signals and velocity signals for the six joints.

8. The control system for the six-joint robot according to claim 4, characterized in that, The position sensor is installed at the angular position measurement point of each joint and is coaxial or rigidly coupled to the joint rotation axis. The velocity sensor is installed at the angular velocity measurement point coaxial with the angular position measurement point to obtain the actual angular velocity of the joint.

9. The control system for the six-joint robot according to claim 1, characterized in that, The layered error includes a position error layer and a velocity error layer, and the position tracking error and velocity tracking error of the six joints are combined in the joint channel order to characterize the current tracking deviation.

10. The control system for the six-joint robot according to claim 5, characterized in that, The residual signal includes at least the difference between the actual joint position and the predicted joint position of each joint, and the difference between the actual joint velocity and the predicted joint velocity. The predicted joint position and the predicted joint velocity are output by the nominal dynamic model under the same control input conditions.

Images