A capture control method and system for a space dual-arm robot based on sliding mode control

Abstract

A method and system for capturing and controlling a space dual-arm robot based on sliding mode control, relating to the field of space dual-arm robot control. It solves the problem of difficult capture and control of dual-arm robots due to the presence of non-cooperative targets in space missions. The method includes: establishing a dynamic model of the space dual-arm robot; predicting the target's motion state and obtaining its trajectory using an unscented Kalman filter algorithm; planning the trajectory of the dynamic model based on the target trajectory; establishing a sliding mode controller based on a switching function; and controlling the dynamic model to move along the planned trajectory according to the sliding mode controller based on the switching function, thereby completing the dual-arm robot capture task. This invention is applied in the aerospace field.

Description

Technical Field

[0001] This invention relates to the field of space dual-arm robot control, and more particularly to a capture control method for a space dual-arm robot based on sliding mode control. Background Technology

[0002] With the continuous development of science and technology and scientists' desire to explore the unknown, on-orbit space operations have become one of the important tasks for humankind in exploring the universe. In space missions, dual-arm robots, as an important tool and technological means, are widely used in spacecraft maintenance, space station construction, satellite on-orbit servicing, and planetary exploration. The flexibility and versatility of space dual-arm robots make them ideal for performing various complex tasks, especially suitable for capturing non-cooperative targets.

[0003] Currently, most space robots still rely on pre-designed mission sequences or real-time command and control by ground operators when performing tasks. However, this traditional control method has certain limitations, especially when facing non-cooperative targets and more complex mission scenarios. In actual space missions, non-cooperative targets are quite common, such as asteroid debris, damaged satellites, or satellite debris. Non-cooperative targets refer to objects that robots need to capture, retrieve, or manipulate, but the motion state of these objects is unknown, and their shape and mass may vary and be uncertain. These unknown motion states and physical properties pose significant challenges to the capture and control of dual-arm space robots. Summary of the Invention

[0004] This invention addresses the difficulty in capturing and controlling dual-armed robots due to the presence of non-cooperative targets in space missions. It proposes a sliding mode control-based method for capturing and controlling a space dual-armed robot. The specific method is as follows:

[0005] A space dual-arm robot capture control method based on sliding mode control, the method comprising:

[0006] Establish a dynamic model of a space dual-arm robot;

[0007] Predict the target's motion state and obtain the target trajectory using the unscented Kalman filter algorithm;

[0008] Based on the target trajectory, perform trajectory planning for the dynamic model;

[0009] Establish a sliding mode controller based on a switching function;

[0010] The sliding mode controller based on the switching function controls the dynamic model to move along the planned trajectory to complete the dual-arm robot capture task.

[0011] Furthermore, a preferred embodiment is provided, wherein the establishment of the dynamic model of the spatial dual-arm robot is as follows:

[0012]

[0013] Among them, H b Floating matrix inertia array, H bm Let H be the inertia matrix of the robotic arm links. m F is the coupling term of the inertia matrix of the matrix and the connecting rod. b For the force acting on the base, F e τ is the external force at the end of the robotic arm. m To apply the control torque to each joint of the robotic arm, c b and c m Together, they constitute the nonlinear term of the robotic arm. For the acceleration of the matrix, J is the angular acceleration of the robotic arm. b J is the velocity Jacobian matrix. w Let be the Jacobian matrix of angular velocity.

[0014] Furthermore, a preferred embodiment is also provided, wherein predicting the target motion state and obtaining the target trajectory based on the unscented Kalman filter algorithm includes:

[0015] The initial values ​​for the first filtering are preset, and the initial values ​​provided by the first filtering include the initial state and the initial covariance matrix;

[0016] Based on the initial state mean and covariance matrix, select 2n+1 sigmapoints for qualitative sampling, and calculate the weight corresponding to each sigmapoint.

[0017] The target motion state is initially predicted using the state transition equation and the weights.

[0018] Calculate the prediction covariance matrix based on the preliminary predicted target motion state;

[0019] Based on the predicted covariance matrix, a second selection of 2n+1 sigmapoints qualitative sampling points is performed, and the weight corresponding to each sigmapoint is calculated.

[0020] Based on the second set of qualitative sampling point weights (sigmapoints), the state propagation of the measurement equation is performed to obtain the predicted state and covariance matrix.

[0021] Error correction is performed based on the predicted state and covariance matrix to obtain the target trajectory.

[0022] Furthermore, a preferred embodiment is also provided, wherein the trajectory planning of the dynamic model based on the target trajectory includes:

[0023] Design the end effector trajectory of the robotic arm in Cartesian space based on the target trajectory;

[0024] Trajectory planning is generated based on the end-effector trajectory and the inverse motion student dynamic model.

[0025] Furthermore, a preferred embodiment is also provided, wherein establishing the sliding mode controller based on the switching function includes:

[0026]

[0027] Where D is the upper bound of external disturbance. This is a hyperbolic tangent switching function. For mass inertia matrix, Here, τ represents the velocity term, and τ represents the control torque. Let c be the desired acceleration, η be the sliding surface coefficient, η be the convergence accuracy, and s be the sliding mode function. This represents the joint angular velocity.

[0028] Furthermore, a preferred embodiment is provided in which the sliding mode controller has a value of 1 for c, a value of 0.5 for η, a value of 0.2 for D, and a value of 0.02 for ε.

[0029] Based on the same inventive concept, the present invention also provides a space dual-arm robot capture and control system based on sliding mode control, the system comprising:

[0030] Modeling unit, used to establish the dynamic model of a space dual-arm robot;

[0031] The prediction unit is used to predict the target's motion state and obtain the target trajectory based on the unscented Kalman filter algorithm.

[0032] A planning unit is used to plan the trajectory of the dynamic model based on the target trajectory;

[0033] The sliding mode controller establishment unit is used to establish a sliding mode controller based on the switching function;

[0034] The control unit is used to control the dynamic model to move along the planned trajectory according to the sliding mode controller based on the switching function.

[0035] Furthermore, a preferred embodiment is provided, wherein the modeling unit is:

[0036]

[0037] Among them, H b Floating matrix inertia array, H bmLet H be the inertia matrix of the robotic arm links. m F is the coupling term of the inertia matrix of the matrix and the connecting rod. b For the force acting on the base, F e τ is the external force at the end of the robotic arm. m To apply the control torque to each joint of the robotic arm, c b and c m Together, they constitute the nonlinear term of the robotic arm. For the acceleration of the matrix, J is the angular acceleration of the robotic arm. b J is the velocity Jacobian matrix. w Let be the Jacobian matrix of angular velocity.

[0038] Based on the same inventive concept, the present invention also provides a computer-readable storage medium for storing a computer program that executes the spatial dual-arm robot capture control method based on sliding mode control described in any of the preceding claims.

[0039] Based on the same inventive concept, the present invention also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and when the processor runs the computer program stored in the memory, the processor executes the spatial dual-arm robot capture control method based on sliding mode control as described in any one of the preceding claims.

[0040] The advantages of this invention are:

[0041] This invention solves the problem of difficulty in capturing and controlling dual-arm robots due to the presence of non-cooperative targets in space missions.

[0042] The sliding mode control-based spatial dual-arm robot capture control method of this invention establishes a dynamic model of the spatial dual-arm robot to accurately describe the robot's dynamic characteristics, including parameters such as mass, inertia, and friction. Based on this model, motion planning and control algorithm design can be performed. The unscented Kalman filter algorithm can more accurately estimate the target's state, exhibiting good convergence and stability. Furthermore, when non-cooperative targets exhibit motion uncertainties and attitude changes, the unscented Kalman filter algorithm can adaptively adjust the model and predictions, providing more accurate target estimation and improving the accuracy and success rate of the capture task. Considering the unpredictability of non-cooperative targets, trajectory planning can continuously update the target trajectory to adapt to changes in the target's attitude and position. In practical applications, a suitable trajectory can be dynamically generated by combining the target's current state and predicted information, enabling the robot to flexibly respond to the motion changes of non-cooperative targets. A sliding mode controller based on a switching function can be established. This switching function can adjust the control strategy in real time according to the target's motion state and the robot's dynamic response to adapt to the uncertainty and changes of non-cooperative targets. Through its fast response and nonlinear control characteristics, the sliding mode controller can track and adjust motion variables such as target attitude and velocity. The robustness and fast response of the sliding mode controller help the robot adapt to the motion and attitude changes of non-cooperative targets. By adjusting the control strategy in real time, the robot can move according to the planned trajectory and remain stable even when the target's motion is uncertain. Through steps such as dynamic modeling, target estimation, trajectory planning, and sliding mode control, the problem of difficult capture and control of dual-arm robots caused by non-cooperative targets in space missions can be effectively solved.

[0043] This invention is applied in the aerospace field. Attached Figure Description

[0044] Figure 1 This is a flowchart of the spatial dual-arm robot capture control method based on sliding mode control as described in Implementation Method 1.

[0045] Figure 2 The flowchart of the UKF algorithm described in Implementation Method 3 is shown below;

[0046] Figure 3 This is a schematic diagram of the space dual-arm robot capturing a target as described in Embodiment Six, wherein, Figure 3 (a) is a schematic diagram of the initial state. Figure 3 (b) is a schematic diagram of a space robot approaching a target and adjusting the pose of its robotic arm. Figure 3 (c) is a schematic diagram of the robotic arm capturing the target. Figure 3 (d) is the target that the robotic arm grabs and retrieves;

[0047] Figure 4This is a schematic diagram of the control forces and torques of the base and robotic arm described in Embodiment Six, wherein, Figure 4 (a) represents the base control force. Figure 4 (b) is the base control torque. Figure 4 (c) represents the control torque of robotic arm #1. Figure 4 (d) represents the control torque of robotic arm #1. Figure 4 (e) represents the control torque of robotic arm #2. Figure 4 (f) represents the control torque of robotic arm #2;

[0048] Figure 5 This is a schematic diagram showing the speed results of the base and robotic arm as described in Embodiment Six, wherein, Figure 5 (a) represents the base velocity. Figure 5 (b) represents the angular velocity of the base. Figure 5 (c) represents the angular velocity of joint 1 of the robotic arm. Figure 5 (d) represents the angular velocity of joint 1 of the robotic arm. Figure 5 (e) represents the angular velocity of the joint of robotic arm #2. Figure 5 (f) represents the angular velocity of the joint of robotic arm No. 2;

[0049] Figure 6 This is a schematic diagram illustrating the control error of the base and robotic arm as described in Embodiment Six, wherein... Figure 6 (a) represents the base position error. Figure 6 (b) represents the base angle error. Figure 6 (c) represents the joint angle error of robotic arm No. 1; Figure 6 (d) represents the joint angle error of robotic arm No. 2;

[0050] Figure 7 This is a schematic diagram of the target motion state as described in Implementation Method Six, wherein, Figure 7 (a) is the target location. Figure 7 (b) is the target angle. Figure 7 (c) represents the target velocity, and (d) represents the target angular velocity. Detailed Implementation

[0051] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0052] Implementation Method 1, see [link] Figure 1 This embodiment describes a spatial dual-arm robot capture and control method based on sliding mode control. The method includes:

[0053] Establish a dynamic model of a space dual-arm robot;

[0054] Predict the target's motion state and obtain the target trajectory using the unscented Kalman filter algorithm;

[0055] Based on the target trajectory, perform trajectory planning for the dynamic model;

[0056] Establish a sliding mode controller based on a switching function;

[0057] The sliding mode controller based on the switching function controls the dynamic model to move along the planned trajectory to complete the dual-arm robot capture task.

[0058] This implementation establishes a dynamic model of a spatial dual-arm robot to accurately describe its dynamic characteristics, including parameters such as mass, inertia, and friction. Based on this model, motion planning and control algorithms can be designed. The unscented Kalman filter algorithm can more accurately estimate the target's state, exhibiting good convergence and stability. Furthermore, when non-cooperative targets exhibit motion uncertainties and attitude changes, the unscented Kalman filter algorithm can adaptively adjust the model and predictions, providing more accurate target estimation and improving the accuracy and success rate of the capture task. Considering the unpredictability of non-cooperative targets, trajectory planning can continuously update the target trajectory to adapt to changes in the target's attitude and position. In practical applications, this can be combined with… The current state and predicted information of the target are used to dynamically generate a suitable trajectory, enabling the robot to flexibly respond to the motion changes of non-cooperative targets. A sliding mode controller based on a switching function is established. The switching function can adjust the control strategy in real time according to the motion state of the target and the dynamic response of the robot to adapt to the uncertainty and changes of non-cooperative targets. Through its fast response and nonlinear control characteristics, the sliding mode controller can track and adjust motion variables such as target posture and velocity. The robustness and fast response characteristics of the sliding mode controller can help the robot adapt to the motion and posture changes of non-cooperative targets. By adjusting the control strategy in real time, the robot can move according to the planned trajectory and remain stable when the target motion is uncertain.

[0059] This implementation method effectively addresses the challenges of capturing and controlling dual-arm robots in space missions caused by non-cooperative targets through steps such as dynamic modeling, target estimation, trajectory planning, and sliding mode control. These steps enable the robot to model, estimate, and predict target motion, and to adopt appropriate control strategies for tracking and capturing, thus addressing the uncertainties and attitude changes of non-cooperative targets.

[0060] Implementation Method Two: This implementation method further defines the spatial dual-arm robot capture control method based on sliding mode control described in Implementation Method One. The establishment of the spatial dual-arm robot dynamic model is as follows:

[0061]

[0062] Among them, H b Floating matrix inertia array, H bm Let H be the inertia matrix of the robotic arm links. m F is the coupling term of the inertia matrix of the matrix and the connecting rod. b For the force acting on the base, F e τ is the external force at the end of the robotic arm. m To apply the control torque to each joint of the robotic arm, c b and c m Together, they constitute the nonlinear term of the robotic arm. For the acceleration of the matrix, J is the angular acceleration of the robotic arm. b J is the velocity Jacobian matrix. w Let be the Jacobian matrix of angular velocity.

[0063] The Lagrange method, starting from the perspective of the system's kinetic and potential energy, expresses the system's equations of motion in a concise and clear form. The core idea of ​​the Lagrange method is to derive the system's equations of motion by analyzing the rate of change of the system's energy.

[0064] This implementation uses the Lagrange method to derive the dynamic model of the space robot, where L represents the Lagrange function, Q is the generalized force vector of the system, T is the kinetic energy, and V is the potential energy, using generalized coordinates. Since the research background is on-orbit operation, the potential energy V is 0.

[0065] Therefore, the Lagrange function is...

[0066] L=T

[0067] Taking the derivative of the equation with respect to time, we can obtain

[0068]

[0069] From the formula, we can obtain

[0070]

[0071] System kinetic energy

[0072]

[0073] Where, m i For the mass of each rod, J i Let the moment of inertia of the connecting rod be _____. Let be the velocity of each rod.

[0074] Substituting the equation, we can obtain the sum of the kinetic energy of the entire system:

[0075]

[0076] In the formula H b H bm H m Defined as follows

[0077]

[0078] For a generalized inertial matrix, H b H is the inertia array of the floating matrix. bm Let H be the inertia matrix of the robotic arm links. m This is the coupling term between the inertia matrices of the base and the connecting rod. Substituting this into the equation yields the general dynamic equation.

[0079]

[0080] In the above formula, let This allows us to derive the generalized dynamic model of the robotic arm.

[0081]

[0082] in, For inertial force, It is a nonlinear term. For Coriolis force, Let it be centrifugal force. c b and c m Together, these constitute the nonlinear term of the robotic arm. Q includes the force F acting on the base. b The external force F at the end of the robotic arm e and the control torque τ acting on each joint of the robotic arm m The dynamic model of the space dual-arm robot is summarized as follows:

[0083]

[0084] Implementation Method 3, see below Figure 2 This embodiment describes a further limitation on the sliding mode control-based spatial dual-arm robot capture control method described in Embodiment 1. The step of predicting the target motion state and obtaining the target trajectory based on the unscented Kalman filter algorithm includes:

[0085] The initial values ​​for the first filtering are preset, and the initial values ​​provided by the first filtering include the initial state and the initial covariance matrix;

[0086] Based on the initial state mean and covariance matrix, select 2n+1 sigmapoints for qualitative sampling, and calculate the weight corresponding to each sigmapoint.

[0087] The target motion state is initially predicted using the state transition equation and the weights.

[0088] Calculate the prediction covariance matrix based on the preliminary predicted target motion state;

[0089] Based on the predicted covariance matrix, a second selection of 2n+1 sigmapoints qualitative sampling points is performed, and the weight corresponding to each sigmapoint is calculated.

[0090] Based on the second set of qualitative sampling point weights (sigmapoints), the state propagation of the measurement equation is performed to obtain the predicted state and covariance matrix.

[0091] Error correction is performed based on the predicted state and covariance matrix to obtain the target trajectory.

[0092] This implementation constructs a non-cooperative target as a rotating target dynamics model, and the rotating target dynamics model is constructed with the following discrete-form state equations and measurement equations:

[0093]

[0094] In the formula, x k The system's state variables [q0, q1, q2, q3, ω1, ω2, ω3] contain seven values, including attitude quaternions and angular velocities; the system noise w is included in the state equation. k It is Gaussian white noise with a mean of 0; the noise v in the measurement equation k This is also Gaussian white noise with a mean of 0. In addition to these quantities, the initial covariance matrix needs to be defined: P0 is the initial covariance matrix, representing the degree of confidence in the initial estimate x0; the covariance matrix Q... k Represents the degree of uncertainty in modeling (errors introduced by the degree of linearization and discretization); covariance matrix R k This represents the degree of uncertainty of the observer.

[0095] The motion state of a rotating target dynamic model is predicted based on an unscented Kalman filter algorithm, wherein the rotating target dynamic model is the target described in this embodiment, and the unscented Kalman filter algorithm includes:

[0096] 1. Initialize the filter

[0097]

[0098] 2. Construct 2n+1 sigma points for the first time and calculate the corresponding weight coefficients.

[0099]

[0100] in, n is the dimension of the state vector. In the formula...

[0101] λ=α 2 (n+k)-n(3-11)

[0102] Where α takes values ​​in

[10] -4 Within the range of ,1]. k=3-n. λ is used to adjust the sigma point and The distance between them. The weights corresponding to the obtained sigma points are...

[0103]

[0104] The value of β in the formula is related to the distribution of the random variable and is used to adjust the accuracy of the covariance.

[0105] 3. Time Update

[0106] Prior state estimation of the state equation at time k:

[0107]

[0108] 4. Construct 2n+1 sigma points for the second time.

[0109] Using the formula Reconstructing the sigma points to prepare for solving the measurement equations.

[0110]

[0111] 5. Measurement Update

[0112] The sigma point propagates in the measurement equation; measurement predictions and covariance matrix are calculated.

[0113]

[0114] 6. Correction of State Equations: State Quantity Estimates and Error Covariance Matrix

[0115]

[0116] This implementation effectively combines prior information and actual observations through flexible sampling and weight calculation, state prediction and covariance estimation, and error correction processes, achieving accurate estimation and prediction of the motion of non-cooperative targets. The unscented Kalman filter algorithm can flexibly adapt to the uncertainty and attitude changes of non-cooperative targets, providing more accurate target estimation and trajectory tracking, thus solving the difficulties in capture control of dual-arm robots.

[0117] Implementation Method Four: This implementation method further defines the spatial dual-arm robot capture control method based on sliding mode control described in Implementation Method One. The step of trajectory planning for the dynamic model based on the target trajectory includes:

[0118] Design the end effector trajectory of the robotic arm in Cartesian space based on the target trajectory;

[0119] Trajectory planning is generated based on the end-effector trajectory and the inverse motion student dynamic model.

[0120] In this embodiment, the purpose of designing the end effector trajectory of the robotic arm in Cartesian space based on the target trajectory is to determine the motion trajectory of the robotic arm's end effector in Cartesian space, that is, to determine the changes in the pose and velocity of the end effector over time. This step includes: determining the pose of the end effector trajectory and determining the velocity of the end effector trajectory;

[0121] Based on the target trajectory, the desired pose of the robotic arm's end effector, including position and orientation, is specified in Cartesian space; based on application requirements and the robotic arm's kinematic parameters, the joint angles of the robotic arm at each moment are calculated using inverse kinematics methods to achieve the motion of the end effector pose.

[0122] After determining the pose of the end effector trajectory, the linear velocity and angular velocity of the robotic arm's end effector can be calculated by differentiating the pose over time. This velocity information can be used to control the movement of the robotic arm, ensuring that the robotic arm can move smoothly and accurately according to the planned trajectory.

[0123] This implementation allows for flexible specification of the end-effector's target pose and velocity by planning the robotic arm trajectory in Cartesian space, enabling it to handle complex task requirements such as precise trajectory tracking, obstacle avoidance, and collaboration. Planning in Cartesian space simplifies complex dynamic problems into more intuitive geometric problems, making the planning process easier to understand and design, and reducing the complexity requirements of inverse kinematics and dynamic models. By directly planning the end-effector trajectory, the position and attitude of the end-effector can be controlled more precisely, improving the accuracy of the target trajectory. Smooth trajectory following can be achieved through appropriate trajectory interpolation and optimization methods.

[0124] In summary, by designing the end-effector trajectory of the robotic arm in the Cartesian space frame based on the target trajectory and combining it with trajectory planning based on the inverse motion student dynamic model, we can provide more optimized and precise control for the robotic arm's motion.

[0125] Implementation Method 5: This implementation method further defines the spatial dual-arm robot capture control method based on sliding mode control described in Implementation Method 1. The establishment of a sliding mode controller based on a switching function includes:

[0126]

[0127] Where D is the upper bound of external disturbance. This is a hyperbolic tangent switching function. For mass inertia matrix, Here, τ represents the velocity term, and τ represents the control torque. Let c be the desired acceleration, η be the sliding surface coefficient, η be the convergence accuracy, and s be the sliding mode function. This represents the joint angular velocity.

[0128] This implementation design uses a sliding mode function, employing a hyperbolic tangent function as the switching function to reduce chattering in sliding mode control. Simulation experiments are used to compare and verify the robustness of the torque capture control based on the sliding mode variable structure computational method.

[0129] The core of sliding mode control is to construct a sliding surface to describe the error between the expected and actual system state. By guiding the system state to jump up and down and dynamically evolve along the sliding surface, the dynamic error of the system converges to zero. Therefore, the control force generated by sliding mode control belongs to a nonlinear control method. Next, a sliding mode controller is constructed based on the dynamic model of a spatial dual-arm robot. First, the joint angles, joint angular velocities, and angular acceleration errors of the robotic arm are defined as follows:

[0130]

[0131] Design the sliding mode function as follows:

[0132]

[0133] The sliding mode controller based on the switching function is designed as follows:

[0134]

[0135] Where D represents the upper bound of external disturbances, and the range of D values ​​needs to be determined according to the magnitude of external disturbances. If the value of D is too small, it will lead to a decrease in the robustness of the controller, and if it is too large, it will cause jitter; η determines the convergence accuracy. The hyperbolic tangent switching function, compared to shn(s), adjusts the control strength according to the error magnitude, which helps reduce the jitter generated in the control process. The second equation can be broken down into two parts, the first half of which... This can be understood as constructing a sliding surface that describes the error between the expected and actual values ​​of the system state; the latter part... The construction of sliding mode dynamics can take many forms. Essentially, sliding mode dynamics provides a nonlinear control force that drives the system state to move rapidly along the sliding surface.

[0136] This implementation method designs and studies a sliding mode controller with better robust performance, completes the capture control of a space dual-arm robot, and reduces jitter and improves control accuracy compared with the calculated torque method.

[0137] Implementation Method Six, see below Figures 3 to 7 This embodiment further defines the spatial dual-arm robot capture control method based on sliding mode control described in Embodiment 5. The sliding mode controller has a value of c = 1, η = 0.5, D = 0.2, and ε = 0.02.

[0138] The process of this embodiment of the space dual-arm robot capturing a rotating target is as follows: Figure 3 As shown, the target is located 4 meters below the space robot and is in a spinning state. The target is a standard square module with a diameter of 0.4m and a spin velocity of . The capture process consists of three steps: First, the base of the space dual-arm robot moves downward by 1m, while the two robotic arms adjust their end-effector poses according to the target module's pose; then, when the target enters the robotic arms' workspace, the two robotic arms slowly approach the target from different sides, and the arms perform a capture by hugging the target; finally, the arms clamp the target to complete the capture and retrieval, maintaining the base's pose throughout the process.

[0139] The driving force and torque on the base of the space dual-arm robot are as follows Figure 4 As shown, the floating base has a maximum control force of 1200N in the Z direction, driving the base to move towards the Z axis; the maximum control torque of the base in the Z direction reaches 580N·m to control the overall posture of the space robot; the joint control force is kept within the range of 4N·m. Since the two robotic arms have symmetrical mechanism configurations and motion planning, their control torques are almost identical. Furthermore, during the entire control process, the control force or torque curves of both the base and the robotic arms exhibit nonlinear characteristics and obvious jitter, which also reflects the nonlinear compensation characteristics of sliding mode control.

[0140] The velocity and angular velocity of the space dual-arm robot base during the entire capture process are as follows: Figure 5 As shown in (a) and (b), the angular velocities of each joint of the two robotic arms are as follows: Figure 5 As shown in (c), (d), (e), and (f), the maximum velocity of the floating base in the Z direction reaches 0.5 m / s; the maximum angular velocity of the base in the Z direction reaches 0.78 rad / s; the angular velocities of the two robotic arm joints exhibit nonlinear characteristics. Similarly, when the target is captured at the end and a collision occurs, the robotic arm will generate a huge disturbance due to the external force, and it will also continuously vibrate during the movement of the target being held.

[0141] The position and angle errors of the space dual-arm robot base during the entire capture process are as follows: Figure 6 As shown in (a) and (b), the joint velocities of the two robotic arms are as follows: Figure 6 As shown in (c) and (d), the maximum positional error of the base is 2 × 10⁻⁶. -4 m, the maximum base angle error is 3×10 m. -4 rad; the error between the two robotic arms remains at 2×10 under no-load conditions. -2 Within the rad range, due to the addition of constraints at the end of the closed-loop system, the maximum joint angle error of the robotic arm after capturing the target reached 0.06 rad.

[0142] The position and angle of the target to be captured during the entire capture process are as follows: Figure 7 As shown in (a) and (b), the target velocity and angular velocity are as follows: Figure 7 As shown in (c) and (d), the target's velocity was zero before being captured. After being captured, as the end effector of the robotic arm moved at a certain speed along the Z-axis, small velocity disturbances were also generated in the X and Y-axis directions. Before being captured, the target's angular velocity along the Z-axis was 0.26 rad / s, and the X and Y-axis velocities were zero. After being captured, the Z-axis angular velocity decreased, and after de-rotation, there was still an angular velocity disturbance of 0.03 rad / s. Finally, the target moved 0.5 m along the negative Z-axis under the clamping of the two robotic arms. During the movement, the attitude was slightly disturbed, but the overall attitude remained relatively stable.

[0143] The results above show that, in the process of capturing non-cooperative targets, the sliding mode controller can reduce errors and limit jitter to a certain extent. Specifically, compared with the torque method based on linearized feedback, the average velocity jitter of the target capture and control method is improved by 10%, the average angular velocity jitter is reduced by 29%, and the frequency of jitter is reduced, which verifies the robustness of the sliding mode controller.

[0144] Implementation Method Seven: The spatial dual-arm robot capture and control system based on sliding mode control described in this implementation method includes:

[0145] Modeling unit, used to establish the dynamic model of a space dual-arm robot;

[0146] The prediction unit is used to predict the target's motion state and obtain the target trajectory based on the unscented Kalman filter algorithm.

[0147] A planning unit is used to plan the trajectory of the dynamic model based on the target trajectory;

[0148] The sliding mode controller establishment unit is used to establish a sliding mode controller based on the switching function;

[0149] The control unit is used to control the dynamic model to move along the planned trajectory according to the sliding mode controller based on the switching function.

[0150] Implementation Method Eight: This implementation method further defines the spatial dual-arm robot capture and control system based on sliding mode control described in Implementation Method Seven. The modeling unit is:

[0151]

[0152] Among them, H b Floating matrix inertia array, H bm Let H be the inertia matrix of the robotic arm links. m F is the coupling term of the inertia matrix of the matrix and the connecting rod. b For the force acting on the base, F e τ is the external force at the end of the robotic arm. m To apply the control torque to each joint of the robotic arm, c b and c m Together, they constitute the nonlinear term of the robotic arm. For the acceleration of the matrix, J is the angular acceleration of the robotic arm. b J is the velocity Jacobian matrix. w Let be the Jacobian matrix of angular velocity.

[0153] Implementation Method Nine: A computer-readable storage medium according to this implementation method, the computer-readable storage medium being used to store a computer program, the computer program executing the spatial dual-arm robot capture control method based on sliding mode control as described in any one of Implementation Methods One to Six.

[0154] Implementation Method 10: A computer device according to this implementation method includes a memory and a processor. The memory stores a computer program. When the processor runs the computer program stored in the memory, the processor executes the spatial dual-arm robot capture control method based on sliding mode control as described in any one of Implementation Methods 1 to 6.

[0155] Although preferred embodiments of this disclosure have been described, those skilled in the art, upon learning the basic inventive concept, can make further changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of this disclosure. Clearly, those skilled in the art can make various alterations and variations to this disclosure without departing from its spirit and scope. Thus, if such modifications and variations fall within the scope of the claims of this disclosure and their equivalents, this disclosure is also intended to include such modifications and variations.

[0156] Those skilled in the art will understand that embodiments of this disclosure can be provided as methods, systems, or computer program products. Therefore, this disclosure can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this disclosure can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0157] The technical solutions provided by the present invention have been described in further detail above with reference to the accompanying drawings in order to highlight their advantages and benefits, and are not intended to limit the present invention. Any modifications, combinations, improvements and equivalent substitutions of the present invention based on the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A space dual-arm robot capture control method based on sliding mode control, characterized in that, The method includes: Establish a dynamic model of a space dual-arm robot; Predict the target's motion state and obtain the target trajectory using the unscented Kalman filter algorithm; Based on the target trajectory, perform trajectory planning for the dynamic model; Establish a sliding mode controller based on a switching function; The sliding mode controller based on the switching function controls the dynamic model to move along the planned trajectory to complete the dual-arm robot capture task; The step of predicting the target motion state and obtaining the target trajectory based on the unscented Kalman filter algorithm includes: The initial values ​​for the first filtering are preset, and the initial values ​​provided by the first filtering include the initial state and the initial covariance matrix; Based on the initial state mean and covariance matrix, select 2n+1 sigma points for qualitative sampling, and calculate the weight corresponding to each sigma point for qualitative sampling. The target motion state is initially predicted using the state transition equation and the weights. Calculate the prediction covariance matrix based on the preliminary predicted target motion state; Based on the predicted covariance matrix, a second selection of 2n+1 sigma points is performed, and the weights corresponding to each sigma point are calculated. The state is propagated based on the qualitative sampling point weights of the second set of sigma points to obtain the predicted state and covariance matrix; Based on the predicted state and covariance matrix, error correction is performed to obtain the target trajectory; The establishment of the sliding mode controller based on the switching function includes: ， in, The upper limit of external interference, This is a hyperbolic tangent switching function. For mass inertia matrix, For the speed term, To control the torque, For the desired acceleration, The sliding surface coefficient, To improve convergence accuracy, For sliding mode function, This represents the joint angular velocity.

2. The spatial dual-arm robot capture control method based on sliding mode control according to claim 1, characterized in that, The established dynamic model of the space dual-arm robot is as follows: ， in, Floating matrix inertia array Here is the inertia matrix of the robotic arm links. This is the coupling term between the inertia matrices of the matrix and the connecting rod. For the force acting on the base, The external force at the end of the robotic arm, To apply the control torque to each joint of the robotic arm, and Together, they constitute the nonlinear term of the robotic arm. For the acceleration of the matrix, The angular acceleration of the robotic arm. For the velocity Jacobian matrix, Let be the Jacobian matrix of angular velocity.

3. The spatial dual-arm robot capture control method based on sliding mode control according to claim 1, characterized in that, The trajectory planning of the dynamic model based on the target trajectory includes: Design the end effector trajectory of the robotic arm in Cartesian space based on the target trajectory; Trajectory planning is generated based on the end-effector trajectory and the inverse motion student dynamic model.

4. The spatial dual-arm robot capture control method based on sliding mode control according to claim 1, characterized in that, The sliding mode controller The value is 1. The value is 0.

5. The value is 0.

2. The value is 0.

02.

5. A spatial dual-arm robot capture and control system based on sliding mode control, characterized in that, The system includes: Modeling unit, used to establish the dynamic model of a space dual-arm robot; The prediction unit is used to predict the target's motion state and obtain the target trajectory based on the unscented Kalman filter algorithm. A planning unit is used to plan the trajectory of the dynamic model based on the target trajectory; The sliding mode controller establishment unit is used to establish a sliding mode controller based on the switching function; A control unit is configured to control the dynamic model to move along a planned trajectory according to the sliding mode controller based on the switching function; The step of predicting the target motion state and obtaining the target trajectory based on the unscented Kalman filter algorithm includes: The initial values ​​for the first filtering are preset, and the initial values ​​provided by the first filtering include the initial state and the initial covariance matrix; Based on the initial state mean and covariance matrix, select 2n+1 sigma points for qualitative sampling, and calculate the weight corresponding to each sigma point for qualitative sampling. The target motion state is initially predicted using the state transition equation and the weights. Calculate the prediction covariance matrix based on the preliminary predicted target motion state; Based on the predicted covariance matrix, a second selection of 2n+1 sigma points is performed, and the weights corresponding to each sigma point are calculated. The state is propagated based on the qualitative sampling point weights of the second set of sigma points to obtain the predicted state and covariance matrix; Based on the predicted state and covariance matrix, error correction is performed to obtain the target trajectory; The establishment of the sliding mode controller based on the switching function includes: ， in, The upper limit of external interference, This is a hyperbolic tangent switching function. For mass inertia matrix, For the speed term, To control the torque, For the desired acceleration, The sliding surface coefficient, To improve convergence accuracy, For sliding mode function, This represents the joint angular velocity.

6. The space dual-arm robot capture and control system based on sliding mode control according to claim 5, characterized in that, The modeling unit is: ， in, Floating matrix inertia array Here is the inertia matrix of the robotic arm links. This is the coupling term between the inertia matrices of the matrix and the connecting rod. For the force acting on the base, The external force at the end of the robotic arm, To apply the control torque to each joint of the robotic arm, and Together, they constitute the nonlinear term of the robotic arm. For the acceleration of the matrix, The angular acceleration of the robotic arm. For the velocity Jacobian matrix, Let be the Jacobian matrix of angular velocity.

7. A computer-readable storage medium, characterized in that, The computer-readable storage medium is used to store a computer program that executes the spatial dual-arm robot capture control method based on sliding mode control as described in any one of claims 1-4.

8. A computer device, characterized in that: It includes a memory and a processor, wherein the memory stores a computer program, and when the processor runs the computer program stored in the memory, the processor executes the spatial dual-arm robot capture control method based on sliding mode control as described in any one of claims 1-4.

Images