Robot teleoperation method and robot

Abstract

The invention provides a robot teleoperation method and a robot. The robot teleoperation method comprises the steps that an acceleration command, an angular acceleration command and an expected attitude quaternion of an end effector are determined based on the current linear acceleration, the current angular acceleration and the current attitude quaternion of a control end; the current position and the current speed fed back by the end effector are obtained, a torque control signal for controlling the position of the end effector is determined based on the acceleration command, the current position and the current speed, and end acceleration feedforward is obtained; determining a torque control signal for controlling the attitude of the end effector based on the angular acceleration command and the expected attitude quaternion to obtain attitude feedforward; and a joint torque instruction of a mechanical arm joint corresponding to the end effector is determined based on the end acceleration feedforward and the posture feedforward, and the joint torque instruction is sent to a corresponding joint driver so that the end effector can synchronously act along with the control end. According to the invention, the problem of attitude lag of the end effector can be avoided, and the tracking error of the end effector is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of robots, and in particular to a robot teleoperation method and a robot. Background Art

[0002] Robot teleoperation is a technology that enables users to control remote machines or systems in real time through interactive devices (also known as handheld operating terminals or mobile control terminals). Its core goal is to synchronously transmit the operator's movements to the remote robot's end-effector. Existing robot teleoperation methods typically employ mapping control based on "zero-order quantities" (i.e., position and attitude). A handheld operating terminal (such as a teach pendant or 3D mouse) acquires the operator's hand's three-dimensional position (x, y, z) and rotational attitude (roll, pitch, yaw) in space. These position / attitude commands are then directly mapped to the robot's end-effector via network transmission, resulting in discrete or continuous position and attitude tracking. However, zero-order mapping control ignores the operator's hand acceleration and angular acceleration information at the user's control terminal, making it unable to respond promptly to large movements at the control terminal. This results in the end-effector losing synchronization with high-speed changes at the control terminal and even experiencing velocity saturation or failing to keep up with the operator's intentions during continuous acceleration. This results in large tracking errors at the end-effector, reducing the robustness of robot control. Summary of the Invention

[0003] In view of this, the purpose of the present invention is to provide a robot remote operation method and a robot, which can predict the operator's intention in advance to accelerate or decelerate during high-dynamic movement at the control end, avoid the problem of posture lag of the end effector, reduce the tracking error of the end effector, and improve the robustness of the robot control.

[0004] In order to achieve the above objectives, the technical solutions adopted in the embodiments of the present invention are as follows:

[0005] In a first aspect, an embodiment of the present invention provides a robot teleoperation method, wherein the robot includes a movable control terminal and a robotic arm end effector, and the robot control method includes:

[0006] Periodically obtaining the current linear acceleration, current angular acceleration, and current posture quaternion corresponding to the control terminal;

[0007] Determining an acceleration command, an angular acceleration command, and an expected attitude quaternion of the end effector based on the current linear acceleration, the current angular acceleration, and the current attitude quaternion of the control end;

[0008] Acquiring the current position and current velocity fed back by the end effector, determining a torque control signal for controlling the position of the end effector based on the acceleration command, the current position, and the current velocity, and obtaining an end acceleration feedforward;

[0009] Determining a torque control signal for controlling the posture of the end effector based on the angular acceleration command and the desired posture quaternion to obtain a posture feedforward;

[0010] The joint torque instruction of the end effector corresponding to the manipulator joint is determined based on the end acceleration feedforward and the posture feedforward, and the joint torque instruction is sent to the corresponding joint driver so that the end effector follows the synchronous movement of the control end.

[0011] Furthermore, an embodiment of the present invention provides a first possible implementation of the first aspect, wherein the step of determining the acceleration command, angular acceleration command, and desired attitude quaternion of the end effector based on the current linear acceleration, current angular acceleration, and current attitude quaternion of the control end includes:

[0012] Converting the current linear acceleration from the control end coordinate system to the robot coordinate system corresponding to the end effector, and determining the acceleration command based on the current linear acceleration after coordinate conversion and the linear acceleration proportional factor;

[0013] Converting the current angular acceleration from the control end coordinate system to the robot coordinate system corresponding to the end effector, and determining the angular acceleration command based on the current angular acceleration after coordinate conversion and the angular acceleration scaling factor;

[0014] The current posture quaternion is converted from the control end coordinate system to the robot coordinate system corresponding to the end effector to obtain the expected posture quaternion.

[0015] Furthermore, an embodiment of the present invention provides a second possible implementation manner of the first aspect, wherein the step of determining the torque control signal for controlling the position of the end effector based on the acceleration command, the current position, and the current speed includes:

[0016] determining a position error of the end effector based on the current position, and determining a velocity error of the end effector based on the current velocity;

[0017] A torque control signal for controlling the end effector to reach a desired end position is determined based on the acceleration command, the position error, and the velocity error.

[0018] Furthermore, an embodiment of the present invention provides a third possible implementation of the first aspect, wherein the calculation formula for the terminal acceleration feedforward is:

[0019] u p =Ma d +K p (pd -p act )+D p (v d -v act )

[0020] Among them, u p is the torque control signal to control the end effector to reach the desired end position, M is the end equivalent mass matrix, a d is the acceleration command, p d is the desired end position, p act is the current position, v d is the desired velocity of the end effector, v act is the current speed, K p is the position feedback gain, D p is the speed feedback gain.

[0021] Furthermore, an embodiment of the present invention provides a fourth possible implementation manner of the first aspect, wherein the step of determining a torque control signal for controlling the posture of the end effector based on the angular acceleration command and the desired posture quaternion includes:

[0022] Acquiring a current posture quaternion of the end effector, and determining a posture error of the end effector based on the current posture quaternion and the expected posture quaternion;

[0023] acquiring a current angular velocity of the end effector, and determining an angular velocity error of the end effector based on the current angular velocity and a desired angular velocity of the end effector;

[0024] A torque control signal for controlling the end effector to achieve the desired posture quaternion is determined based on the angular acceleration command, the posture error, and the angular velocity error.

[0025] Furthermore, an embodiment of the present invention provides a fifth possible implementation of the first aspect, wherein the calculation formula of the posture feedforward is:

[0026] u o =Iα d +K O θ err +D o ω err

[0027] Among them, u o is the torque control signal for controlling the end effector to reach the desired posture quaternion, α d is the angular acceleration command, I is the inertia matrix of the end effector relative to the connected joint, θ err is the attitude error, ω erris the angular velocity error, K O is the attitude feedback gain, D o is the angular velocity feedback gain.

[0028] Furthermore, an embodiment of the present invention provides a sixth possible implementation of the first aspect, wherein the step of determining the joint torque command of the end effector corresponding to the manipulator joint based on the end acceleration feedforward and the posture feedforward includes:

[0029] The terminal acceleration feedforward and the posture feedforward are converted into joint torque instructions of the mechanical arm joint corresponding to the end effector through the Jacobian matrix.

[0030] Furthermore, an embodiment of the present invention provides a seventh possible implementation manner of the first aspect, wherein, after the step of determining the acceleration command based on the current linear acceleration and the linear acceleration scaling factor after coordinate conversion, the robot teleoperation method further includes:

[0031] The current acceleration of the end effector is acquired, and the value of the linear acceleration scale factor is adaptively updated based on the acceleration command and the current acceleration.

[0032] Furthermore, an embodiment of the present invention provides an eighth possible implementation manner of the first aspect, wherein, after the step of determining the angular acceleration command based on the current angular acceleration and the angular acceleration scaling factor after coordinate conversion, the robot teleoperation method further includes:

[0033] The current angular acceleration of the end effector is acquired, and the value of the angular acceleration proportional factor is adaptively updated based on the angular acceleration command and the current angular acceleration.

[0034] In a second aspect, an embodiment of the present invention further provides a robot comprising: a movable manipulation end, a robotic arm end effector, and a controller, wherein the controller comprises a processor and a storage device;

[0035] The storage device stores a computer program, which, when executed by the processor, executes the method according to any one of the first aspects.

[0036] An embodiment of the present invention provides a robot teleoperation method and a robot, wherein the robot includes a movable control terminal and a robotic arm end effector, and the method includes: periodically obtaining a current linear acceleration, a current angular acceleration, and a current posture quaternion corresponding to the control terminal; determining an acceleration command, an angular acceleration command, and a desired posture quaternion of the end effector based on the current linear acceleration, the current angular acceleration, and the current posture quaternion of the control terminal; obtaining a current position and a current velocity fed back by the end effector, and determining a torque control signal for controlling the position of the end effector based on the acceleration command, the current position, and the current velocity, to obtain an end acceleration feedforward; determining a torque control signal for controlling the posture of the end effector based on the angular acceleration command and the desired posture quaternion, to obtain a posture feedforward; determining a joint torque instruction of a robotic arm joint corresponding to the end effector based on the end acceleration feedforward and the posture feedforward, and sending the joint torque instruction to a corresponding joint driver so that the end effector moves synchronously with the control terminal. The present invention collects the current linear acceleration, current angular acceleration and current attitude quaternion corresponding to the control end to calculate the end acceleration feedforward and attitude feedforward when controlling the end effector movement based on the control end, and uses the second-order information of the linear acceleration and angular acceleration of the control end for feedforward compensation, so that the robotic arm can compensate for the large-scale movements of the control end in advance, and then the end effector can predict the operator's intention in advance during the high-dynamic movement of the control end and accelerate or decelerate, avoiding the problem of attitude lag of the end effector, reducing the tracking error of the end effector, and improving the robustness of the robot control.

[0037] Other features and advantages of the embodiments of the present invention will be described in the following description, or some features and advantages can be inferred or determined without doubt from the description, or can be learned by implementing the above-mentioned technologies of the embodiments of the present invention.

[0038] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, preferred embodiments are given below and described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the specific embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0040] Figure 1 A flow chart of a robot teleoperation method provided by an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0041] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions of the present invention will be described below with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0042] At present, traditional robot teleoperation is usually based on the mapping control of "zero-order quantities" (i.e., position and posture). A handheld operation terminal (such as a teach pendant or a 3D mouse) collects the three-dimensional position (x, y, z) and rotational posture (roll, pitch, yaw) of the operator's hand in space, and directly maps these position / posture instructions to the end effector of the robotic arm through network transmission, forming discrete or continuous position and posture tracking.

[0043] Common zero-order control methods include:

[0044] "Direct mapping" method: The position coordinates of the handheld controller are linearly / proportionally scaled and assigned to the target position of the end of the robot arm, and the hand posture rotation angle is mapped to the posture command of the end flange;

[0045] "Incremental mapping" method: The displacement increments Δx, Δy, Δz, Δroll, Δpitch, and Δyaw sampled at each time by the control end are directly used as incremental change commands for the end effector, and accumulated to form a continuous trajectory;

[0046] "Virtual force feedback" method: Based on zero-order position mapping, force / torque sensors or simulated impedance control are used to provide force feedback to the operator, enhancing the interactive immersion.

[0047] However, the zero-order teleoperation method has the following limitations:

[0048] Insufficient dynamic response: Using only zero-order position / attitude signals cannot accurately reflect the acceleration and deceleration process of the operator's hand in high-dynamic scenarios. Especially when performing rapid shaking, slamming or violent displacement, the end effector of the robot arm is prone to motion delay, trajectory lag and jitter due to control link lag, and cannot be synchronized with the operator's movements.

[0049] Poor interactive experience: In high-speed / high-frequency operation scenarios (such as virtual reality games, combat simulations, or battlefield environment training), the operator's hand acceleration and angular acceleration information are critical. Zero-order mapping ignores this part of information, resulting in the robot arm being unable to accurately reproduce the operator's intentions, easily resulting in a "scissors gap" (inconsistency between the operator's movements and the end movements) and a sense of "false inertia", reducing the interactive immersion.

[0050] Limited control accuracy: Using only position / posture as the control variable cannot form feedforward compensation for the end acceleration and angular acceleration, and the end effector of the robot arm will produce large errors when tracking fast motion trajectories; especially in operational tasks that require instantaneous acceleration / deceleration and direction reversal in a short period of time (such as simulating head collisions, boxing strikes, etc.), end tracking deviations can easily lead to operational failures or safety hazards of human-machine collisions.

[0051] Insufficient adaptability and robustness: Most existing zero-order teleoperation methods rely on fixed-proportion mapping or simple PID control, lacking second-order dynamic analysis and adaptive adjustment of the operator's hand motion characteristics. Frequent parameter adjustments are required in different mission scenarios. In the face of network delays, packet loss, or changes in the dynamic parameters of the robotic arm (such as sudden load changes), it is difficult to maintain stable performance.

[0052] Currently, the main application areas of zero-order teleoperation of robots include: medical remote surgery, where some systems introduce force / torque feedback, but still mainly rely on zero-order mapping, which can meet steady-state accuracy requirements in the absence of high-dynamic collision operations; virtual reality control, where some VR robots combine IMU (inertial measurement unit) to provide linear acceleration or angular velocity information in real time, but only as attitude estimation or filtering input, and do not use the second-order acceleration information for actual motion control mapping; industrial teaching and collaboration, which focuses on trajectory reproduction and slow operation scenarios, where zero-order mapping is sufficient to meet the requirements, but lacks precise tracking of second-order quantities in high-dynamic scenarios such as bombardment and collision experiments.

[0053] The existing zero-order mapping control for robot teleoperation still has the following disadvantages:

[0054] Poor dynamic response: Existing zero-order mapping control ignores the operator's hand acceleration and angular acceleration information on the control side. This causes the robot's end effector to lose synchronization when the control side changes at high speeds. It can even lead to speed saturation or failure to keep up with the operator's intention during continuous acceleration.

[0055] Unsuitable for high-dynamic interactions: Existing zero-order mapping control lacks acceleration feedforward and second-order feedback control, making it impossible to respond promptly to the operator's violent hand movements, affecting the interactive experience in combat, gaming, or training scenarios.

[0056] Difficulty in achieving adaptive control: Zero-order control often requires manual selection of appropriate gain or filter parameters and has limited robustness. In situations where network delays fluctuate or the robot arm is subjected to varying loads, real-time parameter adjustments are often required, increasing operational complexity and maintenance costs.

[0057] Limited posture tracking performance: The existing zero-order mapping control only maps the angle signal, which is difficult to capture the angular acceleration of the operator's hand at the control end, resulting in angle lag or oscillation of the end posture in fast flipping scenarios.

[0058] In order to improve the above problems, an embodiment of the present invention provides a robot teleoperation method and a robot. The embodiment of the present invention is introduced in detail below.

[0059] This embodiment provides a robot teleoperation method, which can be applied to a controller of a robot comprising a movable control terminal and a manipulator end effector, see Figure 1 The robot teleoperation method flow chart shown in FIG. 1 mainly includes the following steps:

[0060] Step S102, periodically obtaining the current linear acceleration, current angular acceleration, and current posture quaternion corresponding to the control terminal;

[0061] The control terminal can be a handheld device that integrates a six-axis IMU (inertial measurement unit) and a six-degree-of-freedom attitude sensor to collect the linear acceleration α of the operator's hand in three-dimensional space in real time. h =[ax,ay,az], angular acceleration α h =[α x ,α y ,α z ] and the current posture quaternion, the parameter collection of the control end can be collected using the tactile feedback module. The above posture quaternion q h =[w,x,y,z] is a general hand gesture representation method that satisfies w 2 +x 2 +y 2 +z 2 =1.

[0062] In one embodiment, the control terminal may include a built-in microcontroller to preprocess the collected sensor data (including the current linear acceleration, current angular acceleration, and current attitude quaternion). This preprocessing includes gravity removal, low-pass filtering (such as using a second-order Butterworth low-pass filter to eliminate high-frequency noise), and attitude resolution (updating the attitude quaternion based on the filtered angular velocity). The sampling frequency of the control terminal's sensor data can be greater than or equal to 200Hz to meet the stringent data update rate requirements of high-dynamic scenarios.

[0063] To remove the gravity component, a high-pass filter is used to directly filter out the pure acceleration signal. Considering that the gravity component appears as a DC (zero-frequency) component in acceleration measurement, an extremely low cutoff frequency fc (such as 0.1–0.3 Hz) is set. All components below fc are attenuated by the high-pass filter, which can approximately remove the gravity component.

[0064] The above-mentioned control end and robot body can both include a network communication module, and the control end transmits the current linear acceleration, current angular acceleration and current posture quaternion of the operator's hand collected at the control end to the robot controller through the network communication module, so that the controller controls the end effector of the robotic arm according to the current linear acceleration, current angular acceleration and current posture quaternion of the operator's hand collected by the control end.

[0065] The above-mentioned network communication module can be based on UDP / TCP or real-time ROS2 distributed architecture to ensure that the network transmission delay is ≤20ms and the packet loss rate is ≤1%. At the same time, a timestamp synchronization mechanism is added to ensure the consistency of the data time stamps between the hand and the robot, which facilitates the timing alignment of the second-order quantities.

[0066] Step S104, determining the acceleration command, angular acceleration command and expected attitude quaternion of the end effector based on the current linear acceleration, current angular acceleration and current attitude quaternion of the control end;

[0067] The above-mentioned robot arm has high-speed movement capability and a wide acceleration / deceleration range (end speed ≤ 1m / s, maximum acceleration up to 2m / s 2 ); The robot's controller supports custom control loops, enabling second-order feedforward compensation and adaptive parameter adjustment.

[0068] The controller converts the current linear acceleration, current angular acceleration and current posture quaternion in the control end coordinate system to the robot coordinate system through the coordinate rotation matrix, and generates the acceleration command, angular acceleration command and expected posture quaternion of the end effector based on the linear acceleration, current angular acceleration and current posture quaternion of the control end in the robot coordinate system.

[0069] Step S106, obtaining the current position and current velocity fed back by the end effector, determining a torque control signal for controlling the position of the end effector based on the acceleration command, the current position and the current velocity, and obtaining an end acceleration feedforward;

[0070] The controller's control algorithm for the robot includes a position controller and an attitude controller. The position controller can calculate the position error and velocity error of the end effector based on the current position and current velocity feedback from the end effector, and generate the force or torque control signal of the position controller based on the acceleration command, position error and velocity error. The torque signal is used as the end acceleration feedforward, thereby realizing the hybrid control of second-order feedforward and zero-order feedback.

[0071] Step S108, determining a torque control signal for controlling the attitude of the end effector based on the angular acceleration command and the desired attitude quaternion, and obtaining attitude feedforward;

[0072] The attitude controller determines the force or torque control signal of the attitude controller according to the angular acceleration command and the desired attitude quaternion, and uses the torque signal as attitude feedforward.

[0073] Step S110 , determining the joint torque command of the end effector corresponding to the manipulator joint based on the end acceleration feedforward and the posture feedforward, and sending the joint torque command to the corresponding joint driver so that the end effector follows the synchronous movement of the control end.

[0074] The terminal acceleration feedforward and posture feedforward are converted into joint torque instructions of the end effector corresponding to the robotic arm joint through the Jacobian matrix.

[0075] The terminal acceleration feedforward and attitude feedforward [u p ,u o ] is converted into joint torque command τ through the Jacobian matrix J(q) cmd ,After joint deadweight compensation and friction compensation are performed on the joint torque command, it is output ed to the corresponding joint controller.

[0076]

[0077] The above-mentioned robot remote operation method provided in this embodiment collects the current linear acceleration, current angular acceleration and current posture quaternion corresponding to the control end to calculate the end acceleration feedforward and posture feedforward when controlling the end effector movement based on the control end, and uses the second-order information of the linear acceleration and angular acceleration of the control end for feedforward compensation, so that the robotic arm can compensate for the large-scale movements of the control end in advance, and then enable the end effector to predict the operator's intention in advance during the high-dynamic movement of the control end and accelerate or decelerate, avoiding the problem of posture lag of the end effector, reducing the tracking error of the end effector, and improving the robustness of the robot control.

[0078] In one embodiment, this embodiment provides a specific implementation method for determining the acceleration command, angular acceleration command, and desired attitude quaternion of the end effector based on the current linear acceleration, current angular acceleration, and current attitude quaternion of the control end:

[0079] Convert the current linear acceleration from the control end coordinate system to the robot coordinate system corresponding to the end effector, and determine the acceleration command based on the current linear acceleration after coordinate conversion and the linear acceleration scale factor;

[0080] Convert the current angular acceleration from the control end coordinate system to the robot coordinate system corresponding to the end effector, and determine the angular acceleration command based on the current angular acceleration after coordinate conversion and the angular acceleration scale factor;

[0081] The current posture quaternion is converted from the control end coordinate system to the robot coordinate system corresponding to the end effector to obtain the desired posture quaternion.

[0082] Assume that the control end coordinate system (i.e. the reference coordinate system of the operator's hand) is {H}, the robot end fixed coordinate system is {R}, and the current angular acceleration a is calculated by the hand and robot calibration matrix RRH (the rotation relationship between the hand and the robot base). h Projection into the robot coordinate system:

[0083] a r =RRH·a h

[0084] Current acceleration a in the robot coordinate system r Transverse linear acceleration proportional factor K a After scaling, it is used as the acceleration command a of the end effector of the robotic arm d :

[0085] a d =K a ·a r

[0086] The angular acceleration α of the operator's hand h Transform to the robot coordinate system through the rotation matrix RRH:

[0087] α r =RRH·α h

[0088] The current angular acceleration α in the robot coordinate system r After the angular acceleration scaling factor K α After scaling, it is used as the angular acceleration command α of the end effector of the robotic arm d :

[0089] α d =K α α r

[0090] At the same time, the current hand posture quaternion Q h Convert to the robot coordinate system and get the desired posture quaternion Q in the robot coordinate system d As the zero-order reference, the current hand pose at the start of control is aligned with the target end pose for subsequent closed-loop control. The conversion process is to calculate the transformation matrix in the program. The above matrix RRH is a 3×3 rotation matrix from the controller coordinate system to the robot coordinate system.

[0091] In one embodiment, this embodiment provides a specific implementation of determining a torque control signal for controlling the position of the end effector based on an acceleration command, a current position, and a current speed:

[0092] The position error of the end effector is determined based on the current position, and the speed error of the end effector is determined based on the current speed; the torque control signal for controlling the end effector to reach the desired end position is determined based on the acceleration command, the position error and the speed error.

[0093] In a specific embodiment, the calculation formula of the terminal acceleration feedforward is:

[0094] u p =Ma d +K p (p d -p act )+D p (v d -v act )

[0095] Among them, u p is the torque control signal that controls the end effector to reach the desired end position, M is the end equivalent mass matrix, a d is the acceleration command, p d is the desired end position, p act is the current position, v d is the desired velocity of the end effector, v act is the current speed, K p is the position feedback gain, D p is the speed feedback gain.

[0096] Input linear acceleration feedforward a into the position controller d , combined with the position error and velocity error of the end effector, the force or torque control signal of the position controller is calculated.

[0097] In one embodiment, this embodiment provides a specific implementation of determining a torque control signal for controlling the posture of an end effector based on an angular acceleration command and a desired posture quaternion:

[0098] Obtaining the current posture quaternion of the end effector, and determining the posture error of the end effector based on the current posture quaternion and the expected posture quaternion;

[0099] Obtaining a current angular velocity of the end effector, and determining an angular velocity error of the end effector based on the current angular velocity and a desired angular velocity of the end effector;

[0100] The torque control signal for controlling the end effector to reach the desired attitude quaternion is determined based on the angular acceleration command, attitude error and angular velocity error.

[0101] In a specific embodiment, the calculation formula of the posture feedforward is:

[0102] u o =Iαd +K O θ err +D o ω err

[0103] Among them, u o To control the end effector to achieve the desired posture quaternion torque control signal, α d is the angular acceleration command, I is the inertia matrix of the end effector relative to the connected joint, θ err is the attitude error, ω err is the angular velocity error, K O is the attitude feedback gain, D o is the angular velocity feedback gain.

[0104] θ err =log(Q d Q act -1 )

[0105] ω err =ω d -ω act

[0106] Among them, Q d is the desired posture quaternion of the end effector, Q act is the current posture quaternion of the end effector, v d is the desired angular velocity of the end effector, ω act is the current angular velocity of the end effector.

[0107] Input angular acceleration feedforward α into the attitude controller d , combined with the attitude quaternion error θ of the end effector err and angular velocity error ω err , and calculate the force or torque control signal of the attitude controller.

[0108] In one embodiment, after the step of determining the acceleration command based on the current linear acceleration and the linear acceleration scaling factor after coordinate conversion, the method provided in this embodiment further includes:

[0109] Get the current acceleration of the end effector and adaptively update the linear acceleration scale factor based on the acceleration command and the current acceleration.

[0110] During the human-computer interaction process, the operating force, speed and environmental disturbance of the end effector will continue to change. The least squares algorithm is used to periodically and adaptively adjust the linear acceleration proportional factor K. a , linear acceleration proportional factor K a The update calculation formula is:

[0111] Ka =γ a a ract ·a

[0112] Among them, γ a is the learning rate for updating the linear acceleration scale factor, a ract is the actual acceleration of the end, a is the acceleration error, a=a d -a ract , γ a >0 makes K a Following error converges.

[0113] In one embodiment, after the step of determining the angular acceleration command based on the current angular acceleration after coordinate conversion and the angular acceleration scaling factor, the method provided in this embodiment further includes:

[0114] Get the current angular acceleration of the end effector and adaptively update the value of the angular acceleration scale factor based on the angular acceleration command and the current angular acceleration.

[0115] The angular acceleration scaling factor K is periodically and adaptively adjusted using the least squares algorithm. α , angular acceleration scaling factor K α The update calculation formula is:

[0116] K α =γ α α ract α

[0117] Among them, γ α is the learning rate for updating the angular acceleration scale factor, α ract is the actual angular acceleration of the end, α is the angular acceleration error, α=α d -α ract , γ α >0 makes K α Following error converges.

[0118] In one embodiment, the acceleration command of the end effector can be delayed compensated according to the sampling frequency of the control end and the operating cycle of the robot controller. Assuming that the sampling frequency of the control end is fs = 200Hz and the operating cycle of the robot controller is Tc = 5ms, the total communication delay τ from the hand end to the robot arm is measured by the timestamp mechanism. net , and the expected acceleration signal (acceleration command) is predicted and compensated in the control model of the controller. The acceleration command after delay compensation is:

[0119] a dcomp (t) = a h (t+τ net )

[0120] Similarly, linear interpolation prediction (delay compensation) is performed on the angular acceleration signal. For sudden packet loss or jitter, the second-order signal is estimated and interpolated through the Kalman filter to keep the second-order control signal smooth and continuous, avoiding jitter at the end.

[0121] In one embodiment, the approximate values of the terminal equivalent mass matrix M and the inertia matrix I under different load conditions are obtained offline through least squares fitting, and the dynamic parameters can be slightly adjusted online through extended Kalman filtering (EKF) to ensure the effectiveness of second-order feedforward compensation.

[0122] In one embodiment, a variable control bandwidth can be used to adjust the hand movement intensity based on the hand movement intensity. h and α h Dynamically adjust the feedback gains Kp, Dp, Ko, and Do. When high-dynamic motion is detected, the feedback bandwidth is appropriately increased to reduce tracking error; conversely, the bandwidth is reduced to reduce oscillation during low-dynamic or static motion.

[0123] For example, this embodiment provides a periodic acceleration and deceleration (±1m / s 2 ) tested experimental data, and the experimental results show that when the robot is operated and controlled by the above-mentioned robot remote operation method provided by this embodiment, the tracking error RMS of the end effector is ≈ 0.5 cm, while when the robot is operated and controlled by the zero-order scheme (only position mapping), the tracking error RMS of the end effector is ≥ 2 cm. Therefore, the above-mentioned robot remote operation method provided by this embodiment can effectively reduce the tracking error of the end effector.

[0124] In one embodiment, a seamless connection strategy between the second-order and zero-order signals can also be adopted. When the hand is stationary or at low speed, the zero-order position / posture command is used preferentially to ensure smooth operation in weak dynamic scenarios; when high-dynamic movements are detected, such as when the current linear acceleration or current angular acceleration exceeds the corresponding threshold, it automatically switches to the second-order feedforward dominant mode; ensuring that the signal is continuous and jitter-free when switching between the second-order and zero-order.

[0125] Compared with zero-order (position / posture) teleoperation, the robot teleoperation method provided in this embodiment has the following advantages:

[0126] Faster dynamic response, fully utilizing the second-order information of the control end linear acceleration and angular acceleration for feedforward compensation, enabling the end to predict the operator's intention in advance and accelerate / decelerate during high dynamics, significantly reducing tracking error;

[0127] The interactive experience is better in dynamic scenarios. By tracking the dynamic changes of the hand in real time through second-order quantities, the operator's "sudden movements" (such as boxing and waving) can be perceived in real time. The end of the robotic arm can synchronize according to this dynamic curve, allowing the operator to experience a more realistic interactive effect in virtual fighting or gaming scenes. The "scissors gap" effect is reduced, the synchronization of movements is improved, and the user experience is enhanced. The typical user satisfaction score has increased by ≥30%;

[0128] The adaptive robustness is strong. Based on the online least squares / MRAC adaptive adjustment scale factor, the tracking error caused by terminal load changes and network delay fluctuations is significantly improved. Under the conditions of delay fluctuation of ±20ms and terminal load of ±20%, the terminal trajectory error of the above-mentioned robot teleoperation method provided in this embodiment can still be controlled within 1cm, while the error of traditional fixed-gain solutions can easily reach more than 5cm.

[0129] The posture tracking is more accurate. Through the second-order feedforward of angular acceleration, the robot arm can compensate for the large rotation of the hand in advance, which greatly reduces the posture lag and oscillation. 2 During the angular acceleration test, the end-point attitude error of the robot teleoperation method provided in this embodiment is ≤2°, while the error of the zero-order solution is ≥8°;

[0130] It has a wide range of applications and high engineering efficiency. Simply integrating a six-axis IMU into the handheld controller and implementing the aforementioned filtering and preprocessing algorithms eliminates the need for expensive force / torque sensors to achieve high-dynamic teleoperation in most interactive scenarios. When adapting to heterogeneous robotic arms, only the corresponding Jacobian matrix and dynamic parameters need to be re-acquired, without modifying the core second-order feedforward control framework. Compared to pure zero-order solutions, this approach only requires approximately 20% more work in mapping strategies and parameter tuning, yet delivers a performance improvement of ≥30%, offering high cost-performance.

[0131] The introduction of second-order feedforward can reduce the dependence on velocity integral, integrate more dynamic details into the control quantity, and make the cascade control structure more robust;

[0132] It has strong adaptability to different scenarios and a wide range of applications. It can achieve highly dynamic interactions, such as virtual reality combat, robot competition, and human-machine collaborative assembly. It can also realize remote training, such as military simulation and high-frequency vibration simulation in medical rehabilitation. It can also realize industrial demonstrations, such as high-speed handling and impact test benches, which require the restoration of operator acceleration curves.

[0133] Safety and stability have been improved. Under the combined effect of second-order feedforward and adaptive feedback, the system has the ability to quickly converge to sudden collisions, limit violations or packet loss, avoiding large-scale oscillations or loss of control; the experimental alarm time is ≤10ms, and the emergency stop protection mechanism is reliable.

[0134] Corresponding to the robot teleoperation method provided in the above embodiment, an embodiment of the present invention provides a robot, the robot comprising: a movable manipulation end, a robotic arm end effector, and a controller, the controller comprising a processor and a storage device;

[0135] The storage device stores a computer program, which, when executed by the processor, executes the robot teleoperation method provided in the above embodiment.

[0136] The device provided in this embodiment has the same implementation principle and technical effects as those of the aforementioned embodiments. For the sake of brief description, for matters not mentioned in the device embodiment, reference may be made to the corresponding contents in the aforementioned method embodiment.

[0137] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working process of the system described above can refer to the corresponding process in the aforementioned embodiment and will not be repeated here.

[0138] In addition, in the description of the embodiments of the present invention, unless otherwise expressly specified or limited, the terms "mounted," "connected," and "connected" should be understood in a broad sense. For example, they may refer to fixed connections, detachable connections, or integral connections; they may refer to mechanical connections or electrical connections; they may refer to direct connections or indirect connections through an intermediate medium; and they may refer to internal communication between two components. Those skilled in the art will understand the specific meanings of the above terms in the present invention based on the specific circumstances.

[0139] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0140] In the description of the present invention, it should be noted that the terms "center," "upper," "lower," "left," "right," "vertical," "horizontal," "inner," and "outer," etc., indicating orientations or positional relationships, are based on the orientations or positional relationships shown in the accompanying drawings and are intended solely to facilitate and simplify the description of the present invention. They are not intended to indicate or imply that the devices or components referred to must have, be constructed, or operate in a specific orientation, and therefore should not be construed as limitations on the present invention. Furthermore, the terms "first," "second," and "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0141] Finally, it should be noted that the above-described embodiments are only specific implementations of the present invention, which are used to illustrate the technical solutions of the present invention, rather than to limit them. The scope of protection of the present invention is not limited thereto. Although the present invention has been described in detail with reference to the above-described embodiments, those skilled in the art should understand that any person skilled in the art can modify or easily conceive of changes to the technical solutions described in the above-described embodiments within the technical scope disclosed by the present invention, or replace some of the technical features therein with equivalents. Such modifications, changes, or replacements do not deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should be included in the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be based on the scope of protection of the claims.

Claims

1. A robot teleoperation method, characterized in that: The robot includes a movable control end and a robotic arm end effector, and the robot control method includes: Periodically obtaining the current linear acceleration, current angular acceleration, and current posture quaternion corresponding to the control terminal; Determining an acceleration command, an angular acceleration command, and an expected attitude quaternion of the end effector based on the current linear acceleration, the current angular acceleration, and the current attitude quaternion of the control end; Acquiring the current position and current velocity fed back by the end effector, determining a torque control signal for controlling the position of the end effector based on the acceleration command, the current position, and the current velocity, and obtaining an end acceleration feedforward; Determining a torque control signal for controlling the posture of the end effector based on the angular acceleration command and the desired posture quaternion to obtain a posture feedforward; The joint torque instruction of the end effector corresponding to the manipulator joint is determined based on the end acceleration feedforward and the posture feedforward, and the joint torque instruction is sent to the corresponding joint driver so that the end effector follows the synchronous movement of the control end.

2. The method according to claim 1, characterized in that The step of determining the acceleration command, angular acceleration command and expected attitude quaternion of the end effector based on the current linear acceleration, current angular acceleration and current attitude quaternion of the control end includes: Converting the current linear acceleration from the control end coordinate system to the robot coordinate system corresponding to the end effector, and determining the acceleration command based on the current linear acceleration after coordinate conversion and the linear acceleration proportional factor; Converting the current angular acceleration from the control end coordinate system to the robot coordinate system corresponding to the end effector, and determining the angular acceleration command based on the current angular acceleration after coordinate conversion and the angular acceleration scaling factor; The current posture quaternion is converted from the control end coordinate system to the robot coordinate system corresponding to the end effector to obtain the expected posture quaternion.

3. The method according to claim 1, characterized in that The step of determining a torque control signal for controlling the position of the end effector based on the acceleration command, the current position, and the current speed comprises: determining a position error of the end effector based on the current position, and determining a velocity error of the end effector based on the current velocity; A torque control signal for controlling the end effector to reach a desired end position is determined based on the acceleration command, the position error, and the velocity error.

4. The method according to claim 1, wherein The calculation formula of the terminal acceleration feedforward is: at p =Ma d +K p (p d -p act )+D p (in d -v act ) Among them, u p is the torque control signal to control the end effector to reach the desired end position, M is the end equivalent mass matrix, a d is the acceleration command, p d is the desired end position, p act is the current position, v d is the desired velocity of the end effector, v act is the current speed, K p is the position feedback gain, D p is the speed feedback gain.

5. The method according to claim 1, wherein The step of determining a torque control signal for controlling the posture of the end effector based on the angular acceleration command and the desired posture quaternion comprises: Acquiring a current posture quaternion of the end effector, and determining a posture error of the end effector based on the current posture quaternion and the expected posture quaternion; acquiring a current angular velocity of the end effector, and determining an angular velocity error of the end effector based on the current angular velocity and a desired angular velocity of the end effector; A torque control signal for controlling the end effector to achieve the desired posture quaternion is determined based on the angular acceleration command, the posture error, and the angular velocity error.

6. The method according to claim 5, characterized in that The calculation formula of the posture feedforward is: you o =Iα d +K O i err +D o oh err Among them, u o is the torque control signal for controlling the end effector to reach the desired posture quaternion, α d is the angular acceleration command, I is the inertia matrix of the end effector relative to the connected joint, θ err is the attitude error, ω err is the angular velocity error, K O is the attitude feedback gain, D o is the angular velocity feedback gain.

7. The method according to claim 1, characterized in that The step of determining the joint torque instruction of the end effector corresponding to the manipulator joint based on the end acceleration feedforward and the posture feedforward comprises: The terminal acceleration feedforward and the posture feedforward are converted into joint torque instructions of the mechanical arm joint corresponding to the end effector through the Jacobian matrix.

8. The method according to claim 2, characterized in that After the step of determining the acceleration command based on the current linear acceleration and the linear acceleration proportional factor after coordinate conversion, the robot teleoperation method further includes: The current acceleration of the end effector is acquired, and the value of the linear acceleration scale factor is adaptively updated based on the acceleration command and the current acceleration.

9. The method according to claim 2, characterized in that After the step of determining the angular acceleration command based on the current angular acceleration and the angular acceleration proportional factor after coordinate conversion, the robot teleoperation method further includes: The current angular acceleration of the end effector is acquired, and the value of the angular acceleration proportional factor is adaptively updated based on the angular acceleration command and the current angular acceleration.

10. A robot, characterized in that: include: A movable manipulation end, a robotic arm end effector, and a controller, wherein the controller includes a processor and a storage device; The storage device stores a computer program, which, when executed by the processor, executes the method according to any one of claims 1 to 9.

Images