A high-precision force-position hybrid control system for a teleoperation-based robot

Abstract

The application relates to a high-precision force-position hybrid control system of a robot based on remote operation, which comprises a laser tracker used for acquiring the robot pose of the robot in the laser tracker coordinate system; a laser tracker processing unit used for realizing the coordinate conversion of the laser tracker coordinate system and the robot base coordinate system and realizing the real-time calculation of the pose of the robot end in the operation space; a force feedback handle used for providing XYZ three-direction force feedback, realizing the real-time calculation of the offset information of the handle operation space pose and sending the offset information to the robot controller; and a robot controller used for obtaining control instructions sent to the robot through a force-position hybrid control algorithm. The application separates the independent control of the motion and the contact force of the end effector into two independent decoupling sub-problems of position control and force control, guarantees the high-precision force constant in the vertical direction while improving the absolute position control precision of the robot, and thus guarantees the machining effect of polishing, cutting and other applications.

Description

Technical Field

[0001] This invention belongs to the field of robot motion control, specifically relating to a high-precision force-position hybrid control system for robots based on teleoperation. Background Technology

[0002] Industrial robots have broad application prospects in the field of intelligent manufacturing due to their advantages such as flexible deployment, low cost, and ease of operation. However, compared with machine tools, they are less rigid, have lower absolute positioning accuracy, and are easily subjected to external forces that can cause robot vibration, further affecting the robot's machining accuracy. The absolute error can reach the millimeter level, which seriously limits the application of industrial robots in high-precision machining in fields such as grinding and cutting. At the same time, when robots are dealing with complex perception and a large number of tasks, and need to make rapid decisions and handle extreme situations, remote operation is far superior to intelligent programming.

[0003] Currently, improving the absolute positioning accuracy of industrial robots mainly relies on high-precision measuring devices such as laser trackers to calibrate operational parameters, calculate corrected models of the industrial robot, or directly establish absolute position error and joint error models. Since the data is collected only under local working conditions, the model is not applicable to the global task; when the task changes, data needs to be collected again and the model re-identified. On the other hand, high-precision machining processes such as grinding and cutting not only require high positional accuracy but also high force control accuracy, necessitating the application of force-position hybrid control to execute the required robot trajectory. For example, when an industrial robot cuts along a surface, high-precision force control in the vertical direction and high-precision position control in the cutting direction are required. Combining force feedback handles allows for real-time control of the robot's motion trajectory via manual remote control, enhancing the robot's capabilities in extreme environments.

[0004] Therefore, existing methods for improving the absolute accuracy of robots cannot compensate for errors through online correction, nor can they meet the force control requirements of remote operation of robots. They also struggle to effectively improve the control accuracy, stability, and remote operation capabilities of industrial robots in environments with varying working conditions. Therefore, there is an urgent need for a high-precision force-position hybrid control system for robots based on remote operation. Summary of the Invention

[0005] This invention provides a high-precision force-position hybrid control system for robots based on teleoperation. It includes a robot controller, an industrial robot body, a six-dimensional force sensor, a laser tracker, and a force feedback handle. This invention also provides a high-precision force-position hybrid control method for industrial robots, which can simultaneously improve the absolute position control accuracy of the robot and ensure high-precision force control in the vertical direction.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: a high-precision force-position hybrid control system for robots based on teleoperation, comprising:

[0007] A laser tracker is used to obtain the robot's pose in the laser tracker coordinate system by tracking the spatial position changes of the robot's end effector.

[0008] The laser tracker processing unit is used to realize the coordinate transformation between the laser tracker coordinate system and the robot base coordinate system, and to calculate the pose of the robot end effector in real time within the operating space and send it to the robot controller.

[0009] Force feedback handles are used to track human hand operations and provide force feedback in the XYZ three directions. They calculate the offset information of the handle's spatial pose in real time and send it to the robot controller to realize remote operation control of the robot's end effector.

[0010] The robot controller receives the pose of the robot end effector from the laser tracker processing unit, acquires the force information collected by the six force sensors located at the robot end effector, obtains control commands through a force-position hybrid control algorithm, and sends them to the robot body.

[0011] The force-position hybrid control algorithm is as follows:

[0012] The force information is collected by the six force sensors located at the end of the robot. The force target value is obtained by using the offset information of the handle's spatial posture. The force error is obtained by subtracting the force target value from the force information.

[0013] The position and orientation of the robot end effector sent by the laser tracker processing unit are obtained. The target orientation value is obtained by using the offset information of the robot end effector in the space of the handle. The orientation error is obtained by subtracting the target orientation value from the orientation of the robot end effector obtained by the laser tracker processing unit.

[0014] By integrating force error and pose error into a comprehensive position control, a trajectory planning curve is obtained.

[0015] The force target value is obtained by using the offset information of the spatial pose of the handle, and is calculated using the following formula:

[0016]

[0017] Among them, F d (t) represents the target force value, F d `max` is the maximum value of the set force, and `Pz` is the offset value of the force feedback handle in the Z direction.

[0018] The integrated position control specifically includes:

[0019] Integrated position control P(t) = S·P d(t)+S′·P f (t);

[0020] Where Pd represents the robot's Cartesian position obtained from the pose error through the position control law; Pf represents the robot's Cartesian position obtained from the force error through the force control law.

[0021] Matrix S and matrix S' represent the control mode for each degree of freedom in force-position hybrid control; matrix S is a diagonal matrix with elements on the diagonal being either 1 or 0.

[0022] For position control, the element at position 1 in S corresponds to element 0 in S'.

[0023] For force control, the element at position 0 in S corresponds to element 1 in S'.

[0024] A high-precision force-position hybrid control method for robots based on teleoperation includes the following steps:

[0025] The laser tracker obtains the robot's pose in the laser tracker's coordinate system by tracking the spatial position changes of the robot's end effector.

[0026] The laser tracker processing unit realizes the coordinate transformation between the laser tracker coordinate system and the robot base coordinate system, calculates the pose of the robot end effector in real time within the operating space, and sends it to the robot controller.

[0027] The force feedback handle tracks human hand operation and provides force feedback in three directions (XYZ). It calculates the offset information of the handle's spatial pose in real time and sends it to the robot controller to realize remote operation control of the robot's end effector.

[0028] The robot controller receives the robot end-effector pose from the laser tracker processing unit, acquires the force information collected by the six force sensors located at the robot end-effector, obtains control commands through the force-position hybrid control algorithm, and sends them to the robot body.

[0029] The force-position hybrid control algorithm includes the following steps:

[0030] The force information is collected by the six force sensors located at the end of the robot. The force target value is obtained by using the offset information of the handle's spatial posture. The force error is obtained by subtracting the force target value from the force information.

[0031] The position and orientation of the robot end effector sent by the laser tracker processing unit are obtained. The target orientation value is obtained by using the offset information of the robot end effector in the space of the handle. The orientation error is obtained by subtracting the target orientation value from the orientation of the robot end effector obtained by the laser tracker processing unit.

[0032] By integrating force error and pose error into a comprehensive position control, a trajectory planning curve is obtained.

[0033] The force target value is obtained by using the offset information of the spatial pose of the handle, and is calculated using the following formula:

[0034]

[0035] Among them, F d (t) represents the target force value, F d `max` is the maximum value of the set force, and `Pz` is the offset value of the force feedback handle in the Z direction.

[0036] The integrated position control specifically includes:

[0037] Integrated position control P(t) = S·P d (t)+S′·P f (t);

[0038] Where Pd represents the robot's Cartesian position obtained from the pose error through the position control law; Pf represents the robot's Cartesian position obtained from the force error through the force control law.

[0039] Matrix S and matrix S' represent the control mode for each degree of freedom in force-position hybrid control; matrix S is a diagonal matrix with elements on the diagonal being either 1 or 0.

[0040] For position control, the element at position 1 in S corresponds to element 0 in S'.

[0041] For force control, the element at position 0 in S corresponds to element 1 in S'.

[0042] The present invention has the following advantages due to the adoption of the above technical solutions:

[0043] 1) High-precision control across the entire workspace. Employing an online error compensation control method, real-time control accuracy at the hundred-micron level can be guaranteed without an error-free model. This eliminates the need for extensive prior experiments and model identification, and offers strong adaptability to various tasks.

[0044] 2) High-precision force-position hybrid control. The motion and contact force of the independent control end effector are divided into two independent decoupled sub-problems: position control and force control. The force controller adjusts the robot to ensure that while improving the absolute position control accuracy of the robot, a high-precision force is maintained in the vertical direction, thereby ensuring the processing effect of applications such as grinding and cutting.

[0045] 3) The remote operating system constructs a realistic and reliable human-computer interaction scenario, emphasizing human-computer interaction and featuring high-precision motion tracking, real-time dynamic force feedback, safety, and ease of operation. Attached image description:

[0046] Figure 1 This is a diagram of the device architecture of the present invention;

[0047] Figure 2 Structural diagram of the force-position hybrid control method of the present invention;

[0048] Figure 3 The expected flowchart of the teleoperation control of the present invention;

[0049] Figure 4 Application flowchart of the present invention. Detailed implementation method:

[0050] The present invention will be further described below with reference to the accompanying drawings and specific embodiments.

[0051] Example

[0052] like Figure 1 As shown in the diagram, a device architecture diagram of the present invention includes a robot body, a robot controller, a laser tracker, laser tracker processing software, a six-dimensional force sensor, a force feedback handle, and other parts.

[0053] The robot body (including servo drives) can be a commercial or self-developed 6-DOF industrial robot, which needs to provide a position / torque control interface and a joint data acquisition interface with an acquisition cycle of 1ms.

[0054] The robot controller described above is used to deploy a force-position hybrid control algorithm. It can read the joint data of the robot body and send control commands to the robot body, with a control cycle of less than 4ms.

[0055] The laser tracker described is a high-precision, large-size measuring instrument used in industrial measurement systems. During use, a reflector is installed at the end of the robot body to measure changes in the spatial position of the robot's end effector.

[0056] The laser tracker processing software first performs coordinate transformation between the laser tracker coordinate system and the robot base coordinate system through the equipment calibration process. Then, it calculates the position and attitude of the robot end effector in real time within the operating space and sends it to the robot controller. The calculation cycle is 1ms.

[0057] The aforementioned six-position force sensor is mainly installed at the end of the robot to calculate the force on the end of the robot and send it to the robot controller. The transmission period is 1ms.

[0058] The force feedback handle can accurately track human hand operation and provide force feedback in the XYZ three directions. It can also calculate the offset information of the handle's spatial pose in real time and send it to the robot controller, thereby realizing remote operation control of the robot's end effector.

[0059] like Figure 2 The diagram shows the structure of the control method of this invention. It includes algorithms for force control, position control, inverse kinematics, and robot control. The final control command u(t) acts on the robot body, while force sensors and laser trackers measure the force and pose of the robot's end effector. The purpose of force / position hybrid control is to simultaneously and independently control the motion and contact force of the end effector, which can be divided into two independent decoupled sub-problems: the position control problem and the force control problem. A position controller and a force controller are used respectively to control the movement of the manipulator in Cartesian space.

[0060] The force control law is PI control, and the Cartesian position described in the inverse coordinate system is calculated.

[0061]

[0062] Where, ΔF(t)=F d (t)-F e (t), F d (t) represents the desired contact force applied to the robot by the force feedback handle, F. e (t) represents the contact force collected by the sensor. K E Based on the stiffness of the object, K pf and K tf This is the proportional and integral coefficient matrix. The force controller converts the desired force into a positional offset to drive the robot to achieve constant force control.

[0063] The position control rate is achieved using PD control.

[0064] Where, ΔP(t)=P0(t)-P L P(t), P0(t) is the robot trajectory planning curve provided by the teleoperated handle, P L (t) represents the robot end-effector pose matrix measured in real time by the laser tracker. K pd is the proportionality coefficient matrix, and TD is the differential time constant.

[0065] The integrated position control P(t) = S·P d (t)+S′·P f(t), where matrices S and S' are introduced to determine which control mode—position or force—should be used to control each joint of the Cartesian manipulator. Matrix S is a bit-diagonal matrix, with diagonal elements being 1 and 0. For position control, a position with an element of 1 in S corresponds to an element of 0 in S'; for force control, a position with an element of 0 in S corresponds to an element of 1 in S'. Therefore, matrices S and S' act as an interlock switch, setting the control mode for each degree of freedom in Cartesian space. According to S, position and attitude are always subject to either position control or force control, and the combination of position control and force control is arbitrary. Position tracking error is ignored in the force control component, while the desired control force is tracked.

[0066] The inverse kinematics is the transformation matrix of the robot joints and end effector postures established using the DH model for industrial robots.

[0067] The robot controller described can employ traditional PID control algorithms, or methods such as sliding mode control and LQR control, primarily to achieve servo control of each joint of the robot. This example uses PID control. Where: K pr For example, coefficient T Ir Let T be the integration time constant. Dr P is the differential time constant. rob u(t) is the input of the PID controller, and u(t) is the output of the PID controller.

[0068] like Figure 3 The diagram shown is a flowchart of the teleoperation control of the present invention.

[0069] The specific steps are as follows:

[0070] Step S1: Pressing a button on the remote control handle triggers the remote operation process, proceeding to step S2;

[0071] Step S2: Within the first cycle triggered by the button, the poses of the teleoperated device and the robot are acquired and recorded as the initial poses, where PR consists of the robot position PRinit and attitude RRinit, and PH consists of the teleoperated device position PHinit and attitude RHinit.

[0072] Step S3: Communicate with the robot and start the teleoperation procedure.

[0073] Step S4: If the button remains closed, proceed to the next step; if the button is released, the process ends.

[0074] Step S5: Periodically calculate the force feedback handle's pose offset. In each cycle, obtain the transformation relationship P and R of the teleoperated device's position PHi and attitude RHi relative to the initial poses PHinit and RHinit; transform the teleoperated pose transformation relationship into the robot's corresponding pose P0(t).

[0075] Step S6: Calculate the control force F d (t), Where F d `max` is the maximum value of the set force, and `Pz` is the offset value of the force feedback handle in the Z direction.

[0076] Step S7: Calculate P0(t) and F d (t) is sent to the robot controller to guide the robot's control.

[0077] like Figure 4 The diagram shown is an application flowchart of one example of the present invention.

[0078] The specific steps are as follows:

[0079] Step S1: Initialize the system. The industrial robot body enters the startup state and sets the parameters of each control model.

[0080] Step S2: Connect the laser tracker. After turning on the laser tracker, the target needs to be zeroed to ensure that the laser tracker is in normal working condition and establish a communication connection.

[0081] Step S3: Initiate robot teleoperation. Receive force feedback handle control information P0(t) and F. d (t).

[0082] Step S4: Determine whether the robot teleoperation has ended. If the teleoperation has ended, end the process. If it has not ended, proceed to step S5.

[0083] Step S5: Force sensor data preprocessing. The data acquired by the six-dimensional force sensor needs to be calibrated first due to the influence of the load. It should also be noted that the sensor data contains noise, and in order to ensure the accuracy of force control, the sensor data may also need to be filtered.

[0084] Step S6: Laser tracker data processing. The pose data collected by the laser tracker is relative to the coordinates of the laser tracker itself, and needs to be multiplied by the calibration matrix for transformation.

[0085] Step S7: Force-position hybrid control calculation. According to the control method of the present invention, the comprehensive control rate is calculated.

[0086] Step S8: Control command issuance. Send robot control commands to the robot body for terminal control. Then proceed to step S4.

Claims

1. A high-precision force-position hybrid control system for robots based on teleoperation, characterized in that, include: A laser tracker is used to obtain the robot's pose in the laser tracker coordinate system by tracking the spatial position changes of the robot's end effector. The laser tracker processing unit is used to realize the coordinate transformation between the laser tracker coordinate system and the robot base coordinate system, and to calculate the pose of the robot end effector in real time within the operating space and send it to the robot controller. Force feedback handles are used to track human hand operations and provide force feedback in the XYZ three directions. They calculate the offset information of the handle's spatial pose in real time and send it to the robot controller to realize remote operation control of the robot's end effector. The robot controller receives the pose of the robot end effector from the laser tracker processing unit, acquires the force information collected by the six force sensors located at the robot end effector, obtains control commands through the force-position hybrid control algorithm, and sends them to the robot body. The force-position hybrid control algorithm is as follows: The force information is collected by the six force sensors located at the end of the robot. The force target value is obtained by using the offset information of the handle's spatial posture. The force error is obtained by subtracting the force target value from the force information. The position and orientation of the robot end effector sent by the laser tracker processing unit are obtained. The target orientation value is obtained by using the offset information of the robot end effector in the space of the handle. The orientation error is obtained by subtracting the target orientation value from the orientation of the robot end effector obtained by the laser tracker processing unit. By integrating force error and pose error into a comprehensive position control, a trajectory planning curve is obtained. The integrated position control specifically includes: Integrated Position Control ; Where Pd represents the robot's Cartesian position obtained from the pose error through the position control law; Pf represents the robot's Cartesian position obtained from the force error through the force control law. Matrix S and matrix S' represent the control mode for each degree of freedom in force-position hybrid control; matrix S is a diagonal matrix with elements on the diagonal being either 1 or 0. For position control, the element at position 1 in S corresponds to element 0 in S'. For force control, the element at position 0 in S corresponds to element 1 in S'.

2. The high-precision force-position hybrid control system for robots based on teleoperation according to claim 1, characterized in that, The force target value is obtained by using the offset information of the spatial pose of the handle, and is calculated using the following formula: ； in, For the target value of force, To set the maximum force, Pz is the offset value of the force feedback handle in the Z direction.

3. A high-precision force-position hybrid control method for robots based on teleoperation, characterized in that, Includes the following steps: The laser tracker obtains the robot's pose in the laser tracker's coordinate system by tracking the spatial position changes of the robot's end effector. The laser tracker processing unit realizes the coordinate transformation between the laser tracker coordinate system and the robot base coordinate system, calculates the pose of the robot end effector in real time within the operating space, and sends it to the robot controller. The force feedback handle tracks human hand operation and provides force feedback in three directions (XYZ). It calculates the offset information of the handle's spatial pose in real time and sends it to the robot controller to realize remote operation control of the robot's end effector. The robot controller receives the pose of the robot end effector from the laser tracker processing unit, acquires the force information collected by the six force sensors located at the robot end effector, obtains control commands through the force-position hybrid control algorithm, and sends them to the robot body. The force-position hybrid control algorithm includes the following steps: The force information is collected by the six force sensors located at the end of the robot. The force target value is obtained by using the offset information of the handle's spatial posture. The force error is obtained by subtracting the force target value from the force information. The position and orientation of the robot end effector sent by the laser tracker processing unit are obtained. The target orientation value is obtained by using the offset information of the robot end effector in the space of the handle. The orientation error is obtained by subtracting the target orientation value from the orientation of the robot end effector obtained by the laser tracker processing unit. By integrating force error and pose error into a comprehensive position control, a trajectory planning curve is obtained. The integrated position control specifically includes: Integrated Position Control ; Where Pd represents the robot's Cartesian position obtained from the pose error through the position control law; Pf represents the robot's Cartesian position obtained from the force error through the force control law. Matrix S and matrix S' represent the control mode for each degree of freedom in force-position hybrid control; matrix S is a diagonal matrix with elements on the diagonal being either 1 or 0. For position control, the element at position 1 in S corresponds to element 0 in S'. For force control, the element at position 0 in S corresponds to element 1 in S'.

4. The high-precision force-position hybrid control method for robots based on teleoperation according to claim 3, characterized in that, The force target value is obtained by using the offset information of the spatial pose of the handle, and is calculated using the following formula: ； in, For the target value of force, To set the maximum force, Pz is the offset value of the force feedback handle in the Z direction.

Images