Teleoperated robotic control method and system with adaptive admittance and force feedback

Abstract

The application belongs to the technical field of robots, and discloses a teleoperation robot control method and system with self-adaptive admittance and force feedback, which comprises the following steps: 1. loading preset parameters, and moving a main arm to an initial calibration pose; 2. reading an original operator force vector and an original environment contact force vector to perform digital filtering; 3. calculating a self-adaptive parameter matrix in parallel; 4. solving a current period main arm desired end speed, performing safety limiting processing on components, reading a current actual end pose of the main arm, converting the limited desired speed into a next period main arm target pose, and sending a target joint angle instruction obtained by solving to the main arm; and 5. reading actual joint angles of the main arm in real time and sending the actual joint angles to a slave arm. The method improves the operation intuitiveness, task adaptability and human-computer interaction safety of the system.

Description

Technical Field

[0001] This application belongs to the field of robotics technology, specifically relating to a control method and system for a teleoperated robot with adaptive admittance and force feedback. Background Technology

[0002] Teleoperated robotic systems, especially master-slave systems with force feedback, have wide applications in fields such as medical surgery, space exploration, and disaster relief. Operators control remote slave devices (slave robotic arms) to complete tasks by operating the master device (master robotic arm), while simultaneously sensing the interaction forces between the slave robotic arm and the environment, achieving an "immersive" operation.

[0003] Admittance control is a mainstream technology for achieving force feedback teleoperation. Its basic principle is to establish a virtual mass-spring-damping model, mapping the force applied by the operator to the motion of the robotic arm. In traditional admittance control, key parameters describing this model, such as the virtual damping matrix D_m and the force feedback scaling matrix S, are typically set as fixed constants.

[0004] However, the following technical problems still exist with this fixed parameter setting:

[0005] (1) The contradiction between responsiveness and stability: In order to achieve rapid movement, a small damping is required; however, when performing slow and delicate operations such as precise alignment, a small damping will cause the system to be unstable and susceptible to interference. A fixed damping value cannot simultaneously meet the optimal performance under both working conditions.

[0006] (2) The contradiction between perception and safety: To perceive minute environmental contact forces (such as touching soft objects), a large force feedback scaling factor needs to be set; however, when the slave robotic arm collides with a rigid environment, a large scaling factor will transmit a huge impact force to the operator, which may cause injury. A fixed scaling factor is difficult to balance the fineness of operation and safety.

[0007] (3) Poor task adaptability: For tasks with different attributes (e.g., switching from large-scale rapid transportation to small-scale precision assembly), operators often need to manually readjust control parameters based on experience, which is cumbersome and difficult to achieve optimal results.

[0008] Therefore, a new method is urgently needed to optimize the system's responsiveness, stability, perception fineness, and operational safety by adaptively adjusting control parameters based on the system's real-time status. Summary of the Invention

[0009] To address the aforementioned technical problems in the existing technology, the purpose of this invention is to enable remotely operated robots to provide optimal operational feel and safety assurance at different stages of a task. The technical solution is as follows:

[0010] A control method for a teleoperated robot with adaptive admittance and force feedback includes the following steps:

[0011] Step 1: System Initialization

[0012] After loading the preset parameters, the master and slave robotic arms move to the initial calibration pose.

[0013] Step 2: Data Acquisition and Preprocessing

[0014] The original operator force vector and the original environmental contact force vector are read from the force sensor via a high-speed communication bus, and the original force vector is digitally filtered to obtain a smooth and stable operator force vector and environmental contact force vector.

[0015] Step 3: Adaptive Parameter Dynamic Calculation

[0016] Parallel computation of virtual damping and force feedback scaling adaptive parameter matrices;

[0017] Step 4: Generating core admittance control commands for the master robotic arm

[0018] The expected end-effector velocity vector of the master robot arm in the current cycle is calculated, and each component is subjected to safety limiting. The current actual end-effector pose of the master robot arm is read, and the limited expected velocity is converted into the target pose of the master robot arm in the next cycle. The target joint angle command is obtained through inverse kinematics calculation and the command is sent to the underlying servo driver of the master robot arm.

[0019] Step 5: Generating and tracking robot arm commands

[0020] The actual joint angles of the master robotic arm are read in real time and sent directly to the underlying servo driver of the slave robotic arm as the target joint angle command in the same cycle.

[0021] Furthermore, the preset parameters in step 1 include: minimum and maximum virtual damping coefficients, minimum and maximum force feedback scaling coefficients, gain and threshold of the velocity-damping response curve, gain and threshold of the force scaling response curve, force signal filtering coefficients, and maximum allowable speed.

[0022] Furthermore, in step 2, the controller performs digital filtering on the original force signals using a first-order low-pass filter. The expression for this filtering is as follows:

[0023] F_filtered(k)=α*F_raw(k)+(1-α)*F_filtered(k-1)

[0024] Where α is the filter coefficient, k represents the current period, and k-1 represents the previous period.

[0025] Furthermore, the steps for calculating the adaptive virtual damping D_m(k) in step 3 are as follows:

[0026] (1) The controller takes the expected speed calculated in the previous cycle as an approximate estimate of the current speed;

[0027] (2) D_m(k) is a 6x6 diagonal matrix, and all diagonal elements are assigned the same value;

[0028] (3) The virtual damping value is calculated using an S-shaped function, the expression of which is:

[0029]

[0030] in, d_min represents the expected velocity of the previous cycle, d_max represents the minimum virtual damping coefficient, d_max represents the maximum virtual damping coefficient, k_v represents the gain of the velocity-damping response curve, and v_ref represents the threshold of the velocity-damping response curve.

[0031] Furthermore, the method for calculating the adaptive force feedback scaling S(k) in step 3 is as follows:

[0032] S(k) is a 6x6 diagonal matrix, whose diagonal elements are calculated by adjusting the gain sign of an inverse S-shaped function. This function takes the environmental contact force norm |F_e(k)| of the current period as input, and its calculation expression is:

[0033]

[0034] Where s_min represents the minimum feedback scaling factor, s_max represents the maximum feedback scaling factor, k_f represents the gain of the force-scaling response curve, and f_ref represents the gain of the force-scaling response curve.

[0035] Furthermore, in step 4, the controller calculates the expected end effector velocity vector of the master robotic arm in the current cycle using the core admittance formula, the expression of which is:

[0036]

[0037] in, F_h(k) represents the expected end-effector velocity vector of the master robotic arm in the current cycle, F_e(k) represents the operator force in the current cycle, and F_e(k) represents the environmental contact force in the current cycle.

[0038] Where D_m(k) and S(k) are matrices that change in real time, and their expressions are as follows:

[0039]

[0040] S(k)=g(||F e (k)||);

[0041] The speed safety limiting process is expressed as follows:

[0042]

[0043] in, This indicates the maximum permissible speed.

[0044] Furthermore, in step 4, after the controller reads the current actual end-effector pose of the master-end manipulator, it uses a numerical integration method to convert the limited desired velocity into the target pose of the master-end manipulator for the next cycle. The expression for this is:

[0045]

[0046] Where X_{k+1} is the target position for the next control cycle, X_k is the actual position for the current cycle, and Δt is the control cycle.

[0047] A teleoperated robot control system with adaptive admittance and force feedback is characterized by comprising: a master robotic arm, a slave robotic arm, a first force six-dimensional sensor, a second force six-dimensional sensor, and a controller; wherein the first force six-dimensional sensor is installed at the end of the master robotic arm to measure the operator's intention, and the second force six-dimensional sensor is installed at the end of the slave robotic arm to measure the environmental interaction force; the slave robotic arm at the remote execution end is controlled by operating the master robotic arm; the controller adopts an asymmetric control architecture, implements admittance control for the master robotic arm, implements position tracking control for the slave robotic arm, and internally runs a high-frequency real-time control loop.

[0048] Beneficial effects: (1) Optimized operation experience: By adaptively adjusting virtual damping, the system achieves a smart feel of "smooth when fast and stable when slow". The operator feels smooth when making large-scale movements and stable and reliable when making precise alignment. (2) Enhanced perception: By adaptively adjusting force feedback scaling, the system can amplify weak contact forces, allowing the operator to clearly perceive slight touches, greatly improving the delicacy of operation in fine tasks. (3) Improved safety: In the event of an accidental collision, the system can automatically attenuate the impact force transmitted to the operator, effectively protecting the operator's personal safety and the safety of the main end robotic arm equipment. (4) Wide task adaptability: One system can adapt well to various types of teleoperation tasks, from rapid handling to fine assembly, without reconfiguring parameters, greatly improving the system's versatility and efficiency. Attached Figure Description

[0049] Figure 1 This is a flowchart illustrating the algorithm execution of the overall control system for a teleoperated robot with adaptive admittance and force feedback according to the present invention.

[0050] Figure 2 This is a framework diagram of the adaptive force feedback algorithm of the present invention. Detailed Implementation

[0051] The specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0052] like Figure 1 As shown, a control method for a teleoperated robot with adaptive admittance and force feedback includes the following steps:

[0053] Step 1: System Initialization

[0054] When the system starts, the controller first loads various preset parameters, including: minimum and maximum virtual damping coefficients d_min and d_max; minimum and maximum force feedback scaling coefficients s_min and s_max; gain k_v and threshold v_ref of the velocity-damping response curve; gain k_f and threshold f_ref of the force-scaling response curve; force signal filtering coefficient α; and maximum permissible speed. At the same time, the master and slave robotic arms move to the initial calibration pose.

[0055] Step 2: Data Acquisition and Preprocessing

[0056] The controller reads the original operator force vector from the first force sensor via a high-speed communication bus. Read the raw environmental contact force vector from the second force sensor To eliminate high-frequency noise and signal jitter from the sensor, the controller uses a first-order low-pass filter to digitally filter the two original force vectors, obtaining a smooth and stable operator force vector and environmental contact force vector, expressed as follows:

[0057] F_filtered(k)=α*F_raw(k)+(1-α)*F_filtered(k-1)

[0058] Where α is the filter coefficient, k represents the current period, and k-1 represents the previous period.

[0059] Step 3: Adaptive Parameter Dynamic Calculation

[0060] The controller performs parallel calculations of two adaptive parameter matrices: virtual damping and force feedback scaling. Figure 2 As shown,

[0061] (1) Calculate the adaptive virtual damping parameter matrix D_m(k)

[0062] ① The controller takes the expected speed calculated in the previous cycle as an approximate estimate of the current speed;

[0063] ②D_m(k) is a 6x6 diagonal matrix, with all diagonal elements assigned the same value;

[0064] ③ The virtual damping value is calculated using an S-shaped function, the expression of which is:

[0065]

[0066] in, d_min represents the expected velocity of the previous cycle, d_max represents the minimum virtual damping coefficient, d_max represents the maximum virtual damping coefficient, k_v represents the gain of the velocity-damping response curve, and v_ref represents the threshold of the velocity-damping response curve.

[0067] When the operator moves the main arm quickly When the exponential term approaches infinity, the denominator increases, causing D_m(k) to approach d_min, and the system exhibits low damping and high response characteristics; when the operator performs slow and precise operations... The exponential term approaches zero, causing D_m(k) to approach d_max, and the system exhibits high damping and high stability characteristics.

[0068] (2) Calculate the adaptive force feedback scaling parameter matrix S(k):

[0069] S(k) is also a 6x6 diagonal matrix, whose diagonal elements are calculated by adjusting the gain sign of the inverse S-shaped function. This function takes the environmental contact force norm |F_e(k)| of the current period as input, and its calculation expression is:

[0070]

[0071] Where s_min represents the minimum feedback scaling factor, s_max represents the maximum feedback scaling factor, k_f represents the gain of the force-scaling response curve, f_ref represents the gain of the force-scaling response curve, and |F_e(k)| represents the environmental contact force norm of the current cycle.

[0072] When the contact force between the slave arm and the environment is weak, i.e., |F_e(k)| < f_ref, the exponential term approaches zero and S(k) approaches s_max, amplifying the weak force sensation and transmitting it to the operator to enhance the perception fineness; when a severe collision occurs, i.e., |F_e(k)| > f_ref, the exponential term approaches infinity and S(k) approaches s_min, attenuating the impact force to protect the operator.

[0073] Among them, D_m(k) and S(k) are matrices that change in real time, and their expressions are respectively:

[0074]

[0075] S(k) = g(||F_e(k)||).

[0076] Step 4: Generation of the command for the master manipulator in core admittance control

[0077] The controller substitutes the dynamically calculated D_m(k) and S(k) in Step 3, together with the smoothed force signals F_h(k) and F_e(k) obtained in Step 2, into the core admittance control formula to calculate the desired end velocity vector of the master arm for the current cycle Its calculation expression is:

[0078]

[0079] Among them, represents the desired end velocity vector of the master manipulator for the current cycle, F_h(k) represents the force exerted by the operator for the current cycle, and F_e(k) represents the environmental contact force for the current cycle.

[0080] To prevent the manipulator from overspeed due to the operator's misoperation or sudden change of the force signal, the controller performs safety limit processing on each component of the desired end velocity vector of the master manipulator to obtain Its expression is:

[0081]

[0082] Among them, represents the maximum allowable speed..

[0083] The controller reads the current actual end pose X_k of the master arm and uses the numerical integration method to convert the limited desired velocity into the target pose X_{k + 1} of the master arm for the next cycle, and its expression is:

[0084]

[0085] Among them, X_{k + 1} is the target position for the next control cycle, X_k is the actual position for the current cycle, and Δt is the control cycle.

[0086] The target pose X_{k+1} is calculated using inverse kinematics to obtain a set of target joint angle commands, which are then sent to the underlying servo driver of the main arm.

[0087] Step 5: Generating and tracking robot arm commands

[0088] The controller reads the actual joint angles of the master arm in real time and sends them directly to the slave arm's underlying servo driver as the target joint angle command in the same cycle. This direct mapping method within joint space effectively avoids the singularity problem that may be encountered in Cartesian space control, ensuring motion stability and robustness.

[0089] Through the above-mentioned cyclical high-frequency control, the system behavior is intelligently and adaptively adjusted under different speed and force interaction conditions, thereby providing operators with an unprecedented high-fidelity remote operation experience while ensuring safety.

[0090] The teleoperated robot control system of the present invention with adaptive admittance and force feedback includes: a master robotic arm as the master operating end, a slave robotic arm as the remote execution end, a first six-dimensional force sensor installed at the end of the master arm for measuring the operator's intention, a second six-dimensional force sensor installed at the end of the slave arm for measuring the environmental interaction force, and a controller. The controller internally runs a high-frequency real-time control loop, and within each control cycle, strictly follows... Figure 1 The process shown above involves performing the steps described.

[0091] The above description is merely a preferred embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A control method for a teleoperated robot with adaptive admittance and force feedback, characterized in that, Includes the following steps: Step 1: System Initialization After loading the preset parameters, the master and slave robotic arms move to the initial calibration pose. Step 2: Data Acquisition and Preprocessing The original operator force vector and the original environmental contact force vector are read from the force sensor via a high-speed communication bus, and the original force vector is digitally filtered to obtain a smooth and stable operator force vector and environmental contact force vector. Step 3: Adaptive Parameter Dynamic Calculation Parallel computation of virtual damping and force feedback scaling adaptive parameter matrices; The adaptive virtual damping The calculation steps are as follows: (1) The controller takes the expected speed calculated in the previous cycle as an approximate estimate of the current speed; (2) Given a 6x6 diagonal matrix, assign the same value to all diagonal elements; (3) The virtual damping value is calculated using an S-shaped function, the expression of which is: in, This represents the expected velocity of motion in the previous cycle. This represents the minimum virtual damping coefficient. This represents the maximum virtual damping coefficient. The gain represents the velocity-damping response curve. The threshold representing the velocity-damping response curve; The adaptive force feedback scaling It is a 6x6 diagonal matrix whose diagonal elements are composed of inverses. The type function is calculated by adjusting the gain sign, and the function is based on the environmental contact force norm of the current period. Given the input, its calculation expression is: in, Indicates the minimum feedback scaling factor. Indicates the maximum feedback scaling factor. The gain represents the force-scaling response curve. The threshold representing the force-scaling response curve; Step 4: Generating core admittance control commands for the master robotic arm The expected end-effector velocity vector of the master robotic arm in the current cycle is calculated by solving the core admittance formula, and each component is subjected to safety limiting. The current actual end-effector pose of the master robotic arm is read, and the limited expected velocity is converted into the target pose of the master robotic arm in the next cycle. The target joint angle command is obtained by inverse kinematics calculation and the command is sent to the underlying servo driver of the master robotic arm. The expression for calculating the expected end effector velocity vector of the master-end robotic arm in the current cycle is: in, This represents the expected end effector velocity vector of the master robotic arm in the current cycle. This represents the operator force vector for the current cycle. This represents the environmental contact force vector for the current cycle; The safety limiting processing is expressed as follows: in, Indicates the maximum permissible speed; Step 5: Generating and tracking robot arm commands The actual joint angles of the master robotic arm are read in real time and sent directly to the underlying servo driver of the slave robotic arm as the target joint angle command in the same cycle.

2. The teleoperated robot control method with adaptive admittance and force feedback according to claim 1, characterized in that, The preset parameters in step 1 include: minimum and maximum virtual damping coefficients, minimum and maximum force feedback scaling coefficients, gain and threshold of the velocity-damping response curve, gain and threshold of the force-scaling response curve, force signal filtering coefficients, and maximum allowable speed.

3. The teleoperated robot control method with adaptive admittance and force feedback according to claim 1, characterized in that, In step 2, the controller performs digital filtering on the original force signals using a first-order low-pass filter. The expression for this filtering is as follows: in, These are the filter coefficients. Indicates the current period. The force sensor indicates the first The original force signal acquired in each sampling period, This indicates the previous cycle of the current cycle.

4. The teleoperated robot control method with adaptive admittance and force feedback according to claim 1, characterized in that, In step 4, after the controller reads the current actual end-effector pose of the master-end manipulator, it uses a numerical integration method to convert the limited desired velocity into the target pose of the master-end manipulator for the next cycle. The expression is as follows: in, The target position for the next control cycle. This represents the actual position in the current cycle. To control the cycle.

5. A teleoperated robot control system based on the teleoperated robot control method with adaptive admittance and force feedback as described in claim 1, comprising: The system comprises a master robotic arm, a slave robotic arm, a first force six-dimensional sensor, a second force six-dimensional sensor, and a controller. The first force six-dimensional sensor is installed at the end of the master robotic arm to measure the operator's intent, and the second force six-dimensional sensor is installed at the end of the slave robotic arm to measure environmental interaction forces. The slave robotic arm is controlled remotely by operating the master robotic arm. The controller employs an asymmetric control architecture, implementing admittance control on the master robotic arm and position tracking control on the slave robotic arm, and internally running a high-frequency real-time control loop.

Images