Redundant mechanical arm dynamic obstacle avoidance control method and device and computer readable storage medium

Abstract

The embodiment of the invention provides a dynamic obstacle avoidance control method and device for a redundant mechanical arm and a computer readable storage medium, and the method comprises the steps: obtaining monitoring data, collected and transmitted by a sensor on the redundant mechanical arm, of the redundant mechanical arm in each control period, the monitoring data comprises an initial joint angle of each joint and a point closest to an obstacle; constraint conditions of the redundant mechanical arm in the current control period are constructed, the constraint conditions comprise a target tracking constraint function, a dynamic obstacle avoidance constraint function, a physical constraint function and a self-collision constraint function, and the target tracking constraint function is obtained by correcting an original tracking constraint function through a pre-constructed feedback function and an avoidance vector; a cost function of the redundant mechanical arm is constructed, and the cost function and the constraint conditions are combined to obtain a quadratic programming equation set of the redundant mechanical arm; and solving the quadratic programming equation set to obtain the target joint speed of each joint of the redundant mechanical arm.

Description

Technical Field

[0001] The embodiments of the present invention relate to the technical field of robotic arm control, and in particular, to a dynamic obstacle avoidance control method, device, and computer-readable storage medium for a redundant robotic arm. Background Art

[0002] Currently, with the gradual development of robot technology, the application of robotic arms has expanded from fixed platforms in traditional industrial environments to more complex working environments. When the working environment changes, the robotic arm needs to make decisions to ensure its safety. Therefore, the dynamic obstacle avoidance planning algorithm for robotic arms has emerged.

[0003] There is a current dynamic obstacle avoidance planning algorithm for robotic arms that uses the QP (Quadratic Programming) algorithm. The QP algorithm can solve both the dynamic obstacle avoidance planning problem and the inverse kinematics problem simultaneously, and at the same time, the optimal solution conditions are set to ensure the best result. However, the inventor found during specific implementation that when a dynamic obstacle appears in the end path of the robotic arm, the various constraint conditions in the traditional QP algorithm will conflict, resulting in planning failure and affecting the normal path planning of the robotic arm. Summary of the Invention

[0004] The technical problem to be solved by the embodiments of the present invention is to provide a dynamic obstacle avoidance control method for a redundant robotic arm, which can effectively avoid constraint conflicts and improve the stability of obstacle avoidance planning.

[0005] The further technical problem to be solved by the embodiments of the present invention is to provide a dynamic obstacle avoidance control device for a redundant robotic arm, which can effectively avoid constraint conflicts and improve the stability of obstacle avoidance planning.

[0006] The further technical problem to be solved by the embodiments of the present invention is to provide a computer-readable storage medium, which can effectively avoid constraint conflicts and improve the stability of obstacle avoidance planning.

[0007] To solve the above technical problems, the embodiments of the present invention provide the following technical solution: A dynamic obstacle avoidance control method for a redundant robotic arm, including the following steps: Obtain the monitoring data of the redundant robotic arm in each control cycle collected and transmitted by the sensors on the redundant robotic arm. The monitoring data includes the initial joint angles of each joint of the redundant robotic arm and the closest point between the redundant robotic arm and the obstacle. Construct the constraint conditions of the redundant manipulator in the current control cycle based on the initial joint angles and the nearest point. The constraint conditions include a target tracking constraint function based on the initial joint angles and regarding joint velocities, a dynamic obstacle avoidance constraint function based on the nearest point and regarding joint velocities, a physical constraint function based on the velocity escape method combined with the initial joint angles, and a self-collision constraint function based on the initial joint angles and used to prevent collisions between the joints of the redundant manipulator. The target tracking constraint function is obtained by modifying the original tracking constraint function constructed based on the initial joint angles through a pre-constructed feedback function and an avoidance vector; Construct a cost function of the redundant manipulator regarding the initial joint angles, and combine the cost function and the constraint conditions to obtain a quadratic programming equation set of the redundant manipulator; and Use an RNN network model to solve the quadratic programming equation set to obtain the target joint velocities of each joint of the redundant manipulator in the current control cycle.

[0008] Further, the feedback function is pre-constructed according to the following steps based on the difference between the initial pose and the desired pose of the end effector of the redundant manipulator in the current control cycle: Set the feedback function as , where X represents the current pose of the end effector of the redundant manipulator, represents the desired pose of the end effector of the redundant manipulator, and γ is a gain coefficient; Calculate based on the rotation matrix of the redundant manipulator, and convert the calculated error into a six-dimensional pose in the Cartesian space. Then the final feedback function is expressed as: , where and represent the current pose and the desired pose of the end effector of the redundant manipulator respectively, represents the mapping from the homogeneous transformation matrix to the six-dimensional pose in the Cartesian space.

[0009] Further, the avoidance vector is a vector that modifies the original tracking constraint function after the end effector enters the preset avoidance threshold range of the obstacle and is tangent to the vector between the obstacle and the end effector. The construction process of the avoidance vector specifically includes: Set the avoidance vector , set the reference unit vector of the vector from the nearest point on the obstacle to the end effector as , and the avoidance vector is tangent to the reference unit vector ; Use the reference unit vector Construct a new three-dimensional coordinate system with the Z-axis. The new three-dimensional coordinate system is represented as: , then The plane where it is located coincides with the XY plane of the three-dimensional coordinate system, The direction in the three-dimensional coordinate system is represented as the target unit vector , and the target unit vector is optimized using trigonometric functions , then the target unit vector is represented as ; Set the target unit vector and The rotation matrix for the transformation is R, where the rotation angle in the rotation matrix R is β, and the rotation axis is and perpendicular to the target unit vector and , then the rotation angle β is represented as: ; The rotation axis is represented as: ; Based on Rodrigues' rotation formula, the rotation matrix R is represented as: ; The transformation formula for transforming the target unit vector to is represented as: ; where The magnitude of is related to the distance from the nearest point on the obstacle to the end effector and the speed of the obstacle , then can be represented as: ; where is a positive constant used to adjust the magnitude; Set the unit vector pointing from the end effector of the redundant manipulator to the target to be grasped as , then based on the avoidance principle that is conducive to approaching the target to be grasped, construct a solution function: ; where, solve for the optimal α when the is maximized, and based on the optimal α, construct and obtain the avoidance vector .

[0010] Furthermore, the original tracking constraint function is represented as: , where and respectively represent and the first-order derivatives with respect to time t, representing the velocities of the redundant manipulator in Cartesian space and joint space, represents the Jacobian matrix, and θ represents the initial joint angle; The target tracking constraint function is expressed as: ; The construction process of the dynamic obstacle avoidance constraint function specifically includes: Set the distance from the nearest point on the redundant manipulator to the point on the obstacle as , set the distance threshold interval [d 2 , d 1 of the redundant manipulator from the obstacle, then the obstacle avoidance constraint of the velocity level of the redundant manipulator can be expressed as: , , where e is a scalar and , is a positive constant used to adjust the convergence rate of the obstacle avoidance constraint, , where represents the unit vector from the point on the redundant manipulator to the point on the obstacle, expressed as , represents the velocity of the point on the redundant manipulator; Based on the kinematic principle, is expressed as the product of the Jacobian matrix of the point and the velocities of the joints of the redundant manipulator: , where , and is only related to the joints before the point on the redundant manipulator; and Let , , then the dynamic obstacle avoidance constraint function is expressed as: ; The physical constraint function is expressed as: , where and respectively represent the upper limit and lower limit of the motion angles of the joints of the redundant manipulator, and respectively represent the upper limit and lower limit of the motion angular velocities of the joints of the redundant manipulator; The construction process of the self-collision constraint function specifically includes: Set the controlled point on the redundant manipulator as , and the collision avoidance point as . Then, by controlling the speed of the controlled point to prevent it from colliding with the collision avoidance point , the collision avoidance function is expressed as: . Let , then the self - collision constraint function is expressed as .

[0011] Furthermore, the construction of the cost function of the redundant manipulator with respect to the initial joint angles specifically includes: Based on the initial joint angles of the current control period and its previous control period, construct the minimum - velocity - norm function and the smooth joint velocity function for each joint of the redundant manipulator. The minimum - velocity - norm function is expressed as: ; The smooth joint velocity function is expressed as: ; where T represents the control period, respectively represent the joint velocities of each joint of the redundant manipulator in the previous control period of the current control period; and Combine the minimum - velocity - norm function and the smooth joint velocity function to obtain the cost function, which is specifically expressed as: ; where is a positive constant that adjusts the evaluation index.

[0012] Furthermore, the quadratic programming equation system of the redundant manipulator is expressed as: ; where the dynamic collision - avoidance constraint function and the self - collision constraint function are combined. Let , .

[0013] Furthermore, the specific process of using the RNN network model to solve the quadratic programming equation system to obtain the target joint velocity of each joint of the redundant manipulator in the current control period includes: Based on the quadratic programming equation system, construct the Lagrangian function, which is specifically expressed as: ; Based on the KKT conditions, the optimal solution of the Lagrangian function satisfies the following optimal - solution relationship, which is specifically expressed as: ; where represents the projection operator; Perform calculations based on the optimal solution relational expression to obtain a system of optimal solution equations, specifically expressed as: ; Wherein, is the learning rate that affects the convergence speed, and are state variables that affect target tracking and obstacle avoidance respectively; Construct an RNN network model based on the parallel computing principle to solve the system of optimal solution equations to obtain the target joint speed of each joint of the redundant manipulator in the current control cycle.

[0014] Furthermore, the monitoring data further includes a depth image of the target to be grasped by the redundant manipulator, and the method further includes: Frame the target box of the target to be grasped from the depth image based on a preset target detection algorithm, and calculate the three-dimensional spatial coordinates of the target to be grasped based on the depth information of the depth image; Obtain the current pose of the end effector of the redundant manipulator in the current control cycle; and Solve and correct the desired pose of the end effector based on the current pose, the target box, and the three-dimensional spatial coordinates of the target to be grasped.

[0015] On the other hand, to solve the above further technical problems, an embodiment of the present invention provides the following technical solution: A redundant manipulator dynamic obstacle avoidance control device is connected to the redundant manipulator. The device includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the redundant manipulator dynamic obstacle avoidance control method as described in any one of the above.

[0016] On the one hand, to solve the above further technical problems, an embodiment of the present invention provides the following technical solution: A computer-readable storage medium includes a stored computer program. Wherein, when the computer program runs, it controls the device where the computer-readable storage medium is located to execute the redundant manipulator dynamic obstacle avoidance control method as described in any one of the above.

[0017] After adopting the above technical solutions, the embodiments of the present invention at least have the following beneficial effects: After the embodiments of the present invention obtain the monitoring data of the redundant manipulator collected by the sensors on the redundant manipulator in each control cycle, the constraint conditions of the redundant manipulator in the current control cycle are constructed based on the initial joint angles and the nearest points. Among them, since the target tracking constraint function based on the initial joint angles and regarding the joint velocity is obtained by modifying the original tracking constraint function through a pre-constructed feedback function and avoidance vector, the feedback function can improve the tracking accuracy, accelerate the convergence speed, ensure that the end effector of the redundant manipulator will move towards the posture of the target to be grasped, and avoid static and dynamic obstacles during the execution process; and to avoid conflicts between the two constraint conditions of tracking and obstacle avoidance caused by obstacles in the direct path of the end effector of the redundant manipulator to the target to be grasped, the original tracking constraint function is modified by using the avoidance vector, so as to modify the position tracking constraint in the quadratic programming process, so that the redundant manipulator can effectively avoid obstacles and avoid causing constraint conflicts; further, after constructing the cost function and obtaining the quadratic programming equations of the redundant manipulator, the RNN network model is used to solve the target joint velocities of each joint in the current control cycle, so as to realize the dynamic obstacle avoidance control of the redundant manipulator. Description of the Drawings

[0018] Figure 1 It is a flowchart of the steps of an optional embodiment of the dynamic obstacle avoidance control method for the redundant manipulator of the present invention.

[0019] Figure 2 It is a schematic diagram of the construction principle of the dynamic obstacle avoidance constraint function of an optional embodiment of the dynamic obstacle avoidance control method for the redundant manipulator of the present invention.

[0020] Figure 3 It is a schematic diagram of the construction principle of the avoidance vector of an optional embodiment of the dynamic obstacle avoidance control method for the redundant manipulator of the present invention.

[0021] Figure 4 It is a schematic diagram of the principle of calculating the expected posture of the end effector of an optional embodiment of the dynamic obstacle avoidance control method for the redundant manipulator of the present invention.

[0022] Figure 5 It is a schematic diagram of the simulation environment of the seven-axis frank manipulator of an optional embodiment of the dynamic obstacle avoidance control method for the redundant manipulator of the present invention.

[0023] Figure 6 It is a comparison schematic diagram of the simulation motion trajectory of the seven-axis frank manipulator and the traditional QP algorithm of an optional embodiment of the dynamic obstacle avoidance control method for the redundant manipulator of the present invention.

[0024] Figure 7Schematic diagram of the comparison between the distance from the end effector of a seven-axis Frank manipulator to the obstacle and the target to be grasped and the traditional QP algorithm in an alternative embodiment of the dynamic obstacle avoidance control method for the redundant manipulator of the present invention.

[0025] Figure 8 Angular velocities and angles of each joint of the seven-axis Frank manipulator when the traditional QP algorithm only uses the minimum velocity norm as the cost function.

[0026] Figure 9 Angular velocities and angles of each joint of the seven-axis Frank manipulator in an alternative embodiment of the dynamic obstacle avoidance control method for the redundant manipulator of the present invention.

[0027] Figure 10 Schematic diagram of the desktop environment using a six-axis UR5 manipulator in an alternative embodiment of the dynamic obstacle avoidance control method for the redundant manipulator of the present invention.

[0028] Figure 11 Schematic diagram of the comparison between the motion trajectory of the six-axis UR5 manipulator and the traditional QP algorithm in an alternative embodiment of the dynamic obstacle avoidance control method for the redundant manipulator of the present invention.

[0029] Figure 12 Schematic diagram of the comparison between the distance from the end effector of the six-axis UR5 manipulator to the obstacle and the target to be grasped and the traditional QP algorithm in an alternative embodiment of the dynamic obstacle avoidance control method for the redundant manipulator of the present invention.

[0030] Figure 13 Angular velocities and angles of each joint of the six-axis UR5 manipulator when the traditional QP algorithm only uses the minimum velocity norm as the cost function.

[0031] Figure 14 Angular velocities and angles of each joint of the six-axis UR5 manipulator in an alternative embodiment of the dynamic obstacle avoidance control method for the redundant manipulator of the present invention.

[0032] Figure 15 Principle block diagram of an alternative embodiment of the dynamic obstacle avoidance control device for the redundant manipulator of the present invention.

[0033] Figure 16 Functional module diagram of an alternative embodiment of the dynamic obstacle avoidance control device for the redundant manipulator of the present invention. Detailed implementation manners

[0034] The present application will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the following illustrative embodiments and descriptions are only used to explain the present invention and are not intended to limit the present invention. Moreover, in the case of no conflict, the embodiments in the present application and the features in the embodiments can be combined with each other.

[0035] As Figure 1As shown, an alternative embodiment of the present invention provides a redundant manipulator dynamic obstacle avoidance control method, including the following steps: S1: Obtain the monitoring data of the redundant manipulator 1 in each control cycle collected by the sensors on the redundant manipulator 1 and transmitted, where the monitoring data includes the initial joint angles of each joint of the redundant manipulator 1 and the closest point between the redundant manipulator 1 and the obstacle; S2: Based on the initial joint angles and the closest point, construct the constraint conditions of the redundant manipulator 1 in the current control cycle. The constraint conditions include a target tracking constraint function based on the initial joint angles and regarding joint velocity, a dynamic obstacle avoidance constraint function based on the closest point and regarding joint velocity, a physical constraint function based on the velocity escape method combined with the initial joint angles, and a self-collision constraint function based on the initial joint angles and used to prevent the joints of the redundant manipulator 1 from colliding with each other. The target tracking constraint function is obtained by modifying the original tracking constraint function based on the initial joint angles through a pre-constructed feedback function and avoidance vector; S3: Construct a cost function of the redundant manipulator 1 regarding the initial joint angles, and combine the cost function and the constraint conditions to obtain the quadratic programming equations of the redundant manipulator 1; and S4: Use the RNN network model to solve the quadratic programming equations to obtain the target joint velocities of each joint of the redundant manipulator 1 in the current control cycle.

[0036] After the embodiments of the present invention obtain the monitoring data of the redundant manipulator 1 collected by the sensors on the redundant manipulator 1 and transmitted in each control cycle, by constructing the constraint conditions of the redundant manipulator 1 in the current control cycle based on the initial joint angles and the closest point. Among them, since the target tracking constraint function based on the initial joint angles and regarding joint velocity is obtained by modifying the original tracking constraint function through a pre-constructed feedback function and avoidance vector, the feedback function can improve the tracking accuracy, accelerate the convergence speed, ensure that the end effector of the redundant manipulator 1 will move towards the posture of the target to be grasped, and avoid static and dynamic obstacles during the execution process; and to avoid conflicts between the two constraint conditions of tracking and obstacle avoidance caused by obstacles appearing in the direct path of the end effector of the redundant manipulator 1 to the target to be grasped, the original tracking constraint function is modified by using the avoidance vector, so as to modify the position tracking constraint in the quadratic programming process, so that the redundant manipulator 1 can effectively avoid obstacles and avoid causing constraint conflicts; further, after constructing the cost function and obtaining the quadratic programming equations of the redundant manipulator 1, use the RNN network model to solve the quadratic programming equations to obtain the target joint velocities of each joint in the current control cycle, realizing the dynamic obstacle avoidance control of the redundant manipulator 1.

[0037] In an alternative embodiment of the present invention, the feedback function is pre-constructed based on the difference between the initial pose and the desired pose of the end effector of the redundant manipulator 1 in the current control cycle according to the following steps: Set the feedback function as , where X represents the current pose of the end effector of the redundant manipulator 1, represents the desired pose of the end effector of the redundant manipulator 1, and γ is the gain coefficient; Calculate based on the rotation matrix of the redundant manipulator 1, and convert the calculated error into a six-dimensional pose in the Cartesian space. Then the final feedback function is expressed as: , where and represent the current pose and the desired pose of the end effector of the redundant manipulator 1 respectively, represents the mapping from the homogeneous transformation matrix to the six-dimensional pose in the Cartesian space.

[0038] In this embodiment, in order to achieve higher tracking accuracy, accelerate the convergence speed, and meet the needs of the planner, a feedback link is introduced. Here, X represents the current pose of the end effector of the redundant manipulator 1. Specifically, the equation of its change with time can be expressed as ; In addition, since the conversion formula from quaternion to Euler angle may cause jumps in Euler angles, resulting in sudden changes in attitude errors, the rotation matrix of the redundant manipulator 1 in the embodiment of the present invention is calculated , and the calculated error is re-converted into a six-dimensional pose representation in the Cartesian space.

[0039] In an alternative embodiment of the present invention, the avoidance vector is a vector that is tangent to the vector between the obstacle and the end effector after the end effector enters the preset avoidance threshold range of the obstacle. The construction process of the avoidance vector specifically includes: Set the avoidance vector , set the reference unit vector of the vector from the nearest point on the obstacle to the end effector as , and the avoidance vector is tangent to the reference unit vector ; Use the reference unit vector as the Z-axis to construct a new three-dimensional coordinate system, which is expressed as: , then The plane where it is located coincides with the XY plane of the three-dimensional coordinate system, The direction in the three-dimensional coordinate system is represented as the target unit vector , and optimize the target unit vector using trigonometric functions , the target unit vector is represented as ; Set the target unit vector and The rotation matrix for transformation is R, where the rotation angle in the rotation matrix R is β, and the rotation axis is and perpendicular to the target unit vector and , then the rotation angle β is expressed as: ; The rotation axis is expressed as: ; Based on Rodrigues' rotation formula, the rotation matrix R is expressed as: ; The transformation formula for transforming the target unit vector to is expressed as: ; where The magnitude of is related to the distance from the nearest point on the obstacle to the end effector and the velocity of the obstacle , then can be expressed as: ; where is a positive constant for adjusting the magnitude; Set the unit vector pointing from the end effector of the redundant manipulator to the target to be grasped as , then based on the avoidance principle conducive to approaching the target to be grasped, a solution function is constructed: ; where, solving for the optimal α when the is maximized, and based on the optimal α, the avoidance vector is constructed.

[0040] If an obstacle appears in the direct path to the target to be grasped, the two constraints of tracking and obstacle avoidance conflict, resulting in a situation where the solution fails (the solver cannot find a solution that satisfies the constraints) or the planning fails (the solver can find a solution, but the end of the manipulator will fall into a local loop). Therefore, in this embodiment, referring to Figure 2As shown, when the end effector enters the preset avoidance threshold range of the obstacle, the target to be grasped during tracking is locally modified by constructing an avoidance vector tangent to the vector between the obstacle and the end effector, so as to achieve the purpose of avoiding the obstacle; when the space becomes three-dimensional, the selection space of the avoidance vector changes from a straight line to a plane; to calculate the avoidance vector , the design target unit vector is expressed as , therefore, how to select the corresponding can obtain the optimal ; Since the obtained and are not in the same reference coordinate system and coordinate transformation is required. For this purpose, the rotation matrix R is used to represent the transformation of the target unit vector to ; Based on the above, from the principle of real-time performance, to simplify the calculation, the of the target unit vector is sampled and selected from between , and the best result is selected from them according to the designed criterion to construct ; Since the QP planner makes real-time decisions according to the current environmental conditions in each discrete time period and can adapt to a rapidly changing environment, on this basis, appropriate criteria can be selected as a guide to improve the success rate of task execution. The design of the criteria is relatively diverse. In this embodiment, a more convenient method is finally selected, that is, to select the avoidance direction that is most conducive to approaching the task target; since the smaller the angle between the modified local target direction to be grasped and , the larger the calculated , so finally the with the largest is selected to construct .

[0041] It can be understood that when the end effector enters the preset avoidance threshold range of the obstacle, the avoidance vector is calculated accordingly, otherwise the is equal to zero.

[0042] In an alternative embodiment of the present invention, the original tracking constraint function is expressed as: , where and respectively represent and the first-order derivatives with respect to time t, representing the velocities of the redundant manipulator 1 in the Cartesian space and the joint space, represents the Jacobian matrix, and θ represents the initial joint angle; The target tracking constraint function is expressed as: ; Due to the non - linear characteristics of the redundant manipulator, it is relatively difficult to directly calculate the inverse kinematics from the position level. Therefore, its first - order differential form is adopted. In this embodiment, the original tracking constraint function is constructed by using the tracking constraint at the velocity level. Among them, the Jacobian matrix in the functional formula is the derivative of the non - linear mapping F() with respect to the joint variable and can linearly map the joint velocity in the joint space to the velocity of the end - effector in the Cartesian space.

[0043] The construction process of the dynamic obstacle - avoidance constraint function specifically includes: As Figure 3 shown, set the distance from the nearest point on the redundant manipulator to the point on the obstacle as , and set the distance threshold interval [d 2 , d 1 between the redundant manipulator and the obstacle. Then, the obstacle - avoidance constraint at the velocity level of the redundant manipulator can be expressed as: , , where e is a scalar and , is a positive constant used to adjust the convergence speed of the obstacle - avoidance constraint, , where represents the unit vector from the point on the redundant manipulator to the point on the obstacle, expressed as , represents the velocity of the point on the redundant manipulator; Based on the kinematic principle, is expressed as the product of the Jacobian matrix of the point and the velocities of each joint of the redundant manipulator: , where , and is only related to the joints before the point on the redundant manipulator; and Let , , then the dynamic obstacle - avoidance constraint function is expressed as: ; As Figure 3 shown, the closest distance between the redundant manipulator and the obstacle can be obtained by the closest point on the redundant manipulator and the closest point on the obstacle in the three - dimensional spaceis represented; when the distance between the redundant robotic arm and the obstacle is less than the upper limit value of the influence distance threshold range , which means the algorithm needs to start controlling to keep it always not less than the lower limit value of the distance threshold range where collision occurs , that is ; according to the model parameters of the redundant robotic arm, the Jacobian matrix of any point on the robotic arm can be calculated (for example: the Jacobian matrix of point ). )

[0044] The physical constraint function is expressed as: , where and respectively represent the upper limit value and the lower limit value of the motion angle of each joint of the redundant robotic arm 1, and respectively represent the upper limit value and the lower limit value of the motion angular velocity of each joint of the redundant robotic arm 1; According to the velocity escape method, the motion angle of each joint should not exceed the upper limit value and the lower limit value of the motion angle of each joint of the redundant robotic arm 1, and the motion angular velocity of each joint should not exceed the upper limit value and the lower limit value of the motion angular velocity of each joint of the redundant robotic arm 1.

[0045] The construction process of the self - collision constraint function specifically includes: Set the controlled point on the redundant robotic arm 1 as , and the collision - avoidance point as , then the collision - avoidance function for preventing it from colliding with the collision - avoidance point by controlling the velocity of the controlled point is expressed as: , let , then the self - collision constraint function is expressed as .

[0046] Finally, in addition to ensuring that the robotic arm does not collide with obstacles in the environment, the QP for planning also needs to consider the problem that the robotic arm may collide with itself. It can avoid colliding with the collision - avoidance point by controlling the velocity of the controlled point .

[0047] In an optional embodiment of the present invention, the construction of the cost function of the redundant robotic arm 1 with respect to the initial joint angle specifically includes: Based on the initial joint angles of the current control cycle and its previous control cycle, a minimized velocity norm function and a smooth joint velocity function of each joint of the redundant robotic arm 1 are respectively constructed. The minimized velocity norm function (MVN) is expressed as: ; The smooth joint velocity function (MSV) is expressed as: ; where T represents the control period, respectively represent the joint velocities of each joint of the redundant manipulator 1 in the previous control period of the current control period; and Combining the minimized velocity norm function and the smooth joint velocity function to obtain the cost function, which is specifically expressed as: ; where, is a positive constant for adjusting the evaluation index.

[0048] For a redundant manipulator, due to the existence of redundancy, the QP planner will output infinitely many solutions. To eliminate redundancy and select the best solution, the secondary task can be set as the optimization of some performance indicators, which do not belong to the tasks that must be achieved, but can further optimize the trajectory generated by the planner; for this reason, from the actual perspective of motion control, minimizing the joint velocity norm is selected as one of the performance indicators. At the same time, in order to smooth the joint velocity, generate a smoother trajectory curve, and reduce the sudden change of the driving torque of the manipulator, this embodiment adds the index of smooth joint velocity. Different joint velocity constraints at different speed levels are used to meet the requirements, and the cost function is formed by minimization.

[0049] In an alternative embodiment of the present invention, the quadratic programming equation system of the redundant manipulator 1 is expressed as: ; where the dynamic obstacle avoidance constraint function and the self-collision constraint function are combined, and let , .

[0050] In this embodiment, the cost function is formed by minimization, and at the same time, the target tracking constraint function, the dynamic obstacle avoidance constraint function, the physical constraint function, and the self-collision constraint function need to be satisfied in sequence to ensure accurate control of dynamic obstacle avoidance.

[0051] In an alternative embodiment of the present invention, the specific steps of using the RNN network model to solve the quadratic programming equation system to obtain the target joint velocity of each joint of the redundant manipulator 1 in the current control period include: Constructing a Lagrangian function based on the quadratic programming equation system, which is specifically expressed as: ; Based on the KKT conditions, the optimal solution of the Lagrangian function satisfies the following optimal solution relationship, which is specifically expressed as: ; Among them, represents a projection operator; Based on the optimal solution relation, calculations are performed to obtain a system of optimal solution equations, specifically expressed as: ; Among them, is the learning rate that affects the convergence speed, and are state variables that affect target tracking and obstacle avoidance respectively; Based on the parallel computing principle, an RNN network model is constructed to solve the system of optimal solution equations to obtain the target joint velocities of each joint of the redundant manipulator 1 within the current control period.

[0052] In this embodiment, an RNN network model is constructed based on the parallel computing principle to solve the system of optimal solution equations. This solution method can calculate the change of joint velocities in real time within each control period, thereby achieving the purpose of dynamic response.

[0053] In a specific implementation, it is possible that the general expression of the projection operator is: ; Its meaning is to input x, find the y closest to x from the given domain and use it as the return value. This is equivalent to making a restriction on the solution of the angular velocity under the joint constraints, ensuring that the calculated angular velocity does not exceed the physical constraints.

[0054] In an alternative embodiment of the present invention, the monitoring data further includes the depth image of the target to be grasped by the redundant manipulator 1, and the method further includes: Based on a preset target detection algorithm, a target box of the target to be grasped is framed out from the depth image, and the spatial three-dimensional coordinates of the target to be grasped are calculated based on the depth information of the depth image; Obtain the current pose of the end effector of the redundant manipulator 1 in the current control period; and Based on the current pose, the target box, and the spatial three-dimensional coordinates of the target to be grasped, solve and correct the desired pose of the end effector.

[0055] From the perspective of a dynamic environment, the optimal posture required in each step of single-step planning will obviously change. Therefore, for a redundant manipulator, as it approaches the target, it is necessary to determine the grasping posture according to its current state and the state of the task target. Through the object detection algorithm, the target box of the task target can be obtained, and then its spatial coordinates can be further obtained using depth information. At the same time, the manipulator will also publish its own state in real time. Based on this, the embodiment of the present invention proposes a method for calculating the desired posture in each control cycle. Because of the simplicity of its calculation process, it does not occupy computing resources and does not increase the computational cost of the manipulator's dynamic obstacle avoidance planning scheme. Moreover, since the manipulator can change the grasping posture at any time based on this, the flexibility and grasping success rate of the manipulator are greatly increased.

[0056] Specifically, as Figure 4 shown, by selecting the X-axis and Z-axis in the rotation matrix of the current posture of the end effector of the redundant manipulator to transform the appropriate grasping posture, the end posture of the manipulator can be represented by the rotation matrix, that is

[0057] wherein, the Z-axis is set to , representing the direction from the midpoint of the end of the manipulator to the midpoint of the grasping target, and the Y is set to the direction parallel to two sides of the target box, that is , and is obtained through the cross product of and .

[0058] Specifically, to verify the control effect of the embodiment of the present invention, a comparative experiment between the embodiment of the present invention and the traditional QP is carried out using a seven-axis manipulator in the simulation environment and a six-axis manipulator in the real environment, including the dynamic obstacle avoidance effect, the planning time, and the planned trajectory.

[0059] First, an experiment is carried out on a seven-axis frank manipulator in the simulation environment: In these experiments, it is assumed that the QP planner has complete state information of the environment, and the following settings are used for the parameters of the planner: The upper and lower limit values of the angle and angular velocity of each joint of the redundant manipulator are set to: ; ; , ; In the distance threshold interval, m, m; the preset avoidance threshold range for the end effector to enter the obstacle m, gain , the positive constant for adjusting the evaluation index is , adjust the positive constant of the size , the positive constant for adjusting the convergence rate of the obstacle avoidance constraint , learning rate , the initial joint angles of each joint of the redundant manipulator ; In this simulation experiment, the redundant manipulator needs to perform planning in a desktop environment with obstacles to verify the effectiveness of the algorithm. As Figure 5 shown, the manipulator needs to reach the target coordinates in a cluttered desktop environment. At the same time, there is a small ball with a diameter of 5 cm on the line connecting the starting position and the target as an obstacle on the trajectory to verify the effectiveness of the avoidance vector. The trajectory of the small ball is , the position of the target to be grasped is set to , the position of the small ball obstacle is set to . To simplify the calculation, the midpoint of the small ball in the path is used as its geometric abstraction.

[0060] The trajectory result of the redundant manipulator in three-dimensional space, as Figure 6 shown, the red small ball is the obstacle in the path, the magenta line is the abstraction of the seven-axis redundant manipulator. The left figure is the traditional QP algorithm, and the right figure is the tracking method provided by the embodiment of the present invention. Obviously, in the left figure, when an obstacle appears in the path, the tracking and obstacle avoidance constraints of QP will conflict, resulting in the end of the redundant manipulator falling into a local extreme point, thus causing the planning to fail. However, in the right figure, after adding the avoidance vector, when an obstacle enters the avoidance threshold, the avoidance part takes effect, and the tracking task of the QP planner is temporarily changed to an avoidance task. At the same time, due to the established evaluation criterion, the temporarily generated avoidance task will drive the redundant manipulator to move towards the target position as much as possible. Finally, it successfully avoids the obstacle in the path and reaches the target point, and the error between the actual result and the task target can reach within 0.0001 m.

[0061] Further, as shown in Figure 7 , the figure records the distance between the end effector of the redundant manipulator of the traditional QP algorithm and the target to be grasped and the distance between the redundant manipulator and the obstacles in the environment in the embodiment of the present invention. From the data comparison, first, from the data record of the traditional QP algorithm, it can be seen that starting from the 2nd second, due to the appearance of the obstacle small ball in the trajectory of the end of the redundant manipulator, the redundant manipulator moves at the local extreme point, and at the same time, the distance between the redundant manipulator and the obstacle is also much smaller than the set collision distance m. However, after adopting the embodiment of the present invention, the task can be successfully executed and the redundant manipulator can be kept at a safe distance from the obstacle at all times.

[0062] Furthermore, to verify the cost function adopted in the embodiments of the present invention, the traditional method that only uses the minimum velocity norm (MVN) is used as a control, as Figure 8 It can be seen from the left joint velocity diagram that only using the minimum velocity norm results in a smaller velocity range planned by the QP planner and greater fluctuations. However, as Figure 9 shown in the left joint velocity diagram, the velocity range planned by using the cost function proposed in the embodiments of the present invention is slightly larger than that of the traditional algorithm, but still within the physical constraints set in the simulation experiment. At the same time, smoother velocity and angle changes are generated.

[0063] Secondly, experiments are carried out using a six-axis UR5 robotic arm in a real environment: The following settings are used for the parameters of the planner: The upper and lower limit values of the angle and angular velocity of each joint of the redundant robotic arm are set as: , , , ; In the distance threshold interval, , , when the end effector enters the preset avoidance threshold range of the obstacle , gain , a positive constant used to adjust the convergence speed of the obstacle avoidance constraint , adjust size of the positive constant , a positive constant used to adjust the convergence speed of the obstacle avoidance constraint , learning rate , the initial joint angles of each joint of the redundant robotic arm .

[0064] In this experiment, the desktop environment is as Figure 10 shown, where Obstacle1 - Obstacle3 represent the obstacles in the path. After receiving the target information published by the preset node, the robotic arm will move towards the specified pose. To avoid repeating the experiment, only the dynamic obstacle avoidance experiment is carried out in this section. After publishing the target to be grasped, the obstacle is manually moved into the path of the robotic arm to verify the dynamic obstacle avoidance scheme of the control method according to the embodiments of the present invention; Refer to Figure 11 shown, where * represents the sampling points of the closest points of the obstacle at each state during the movement of the robotic arm. The improved QP scheme can successfully avoid the obstacles in the path and reach the target location. And from Figure 11 it can be seen that the end - effector pose grasping strategy proposed in this chapter calculates the optimal grasping pose in real - time after entering a certain distance near the target to be grasped. However, as Figure 12As shown in the figure, when the collision distance is modified to 0.05m, the control method provided by the embodiments of the present invention can still achieve the purpose of obstacle avoidance.

[0065] And from Figure 13 and Figure 14 it can be seen that the weighted smoothing produces a smoother velocity curve than MVN. At the same time, since the set velocity threshold is 0.1, it can be seen that in the field of lower speeds, the joint velocity generated by the cost function provided by the embodiments of the present invention also has a function no worse than MVN, and the generated velocity curve is smoother.

[0066] Finally, the embodiments of the present invention also use whether to add the avoidance vector as a variable to calculate the average planning time of the QP planner under a six-axis and seven-axis robotic arm (simulation). After recording multiple experiments, the average planning time of the algorithm is shown in the following table. It can be seen that the addition of the avoidance vector does not cause a large computational burden on the planner, and the average planning time of the QP planner can well meet the real-time requirements.

[0067]

[0068] In specific implementation, the dynamic obstacle avoidance algorithm proposed in this paper can be used in a robotic arm grasping system. The robotic arm grasping system is composed of a UR5 robotic arm, a Microsoft Azure Kinect DK camera, and a Robotiq 2f-85 two-finger gripper. The software platform is developed based on ROS1 under ubuntu18.04; this system can perform grasping in a desktop environment with static and dynamic obstacles. After specific experiments, different obstacle environments and the positions of the objects to be grasped are selected in the experiment to detect the success rate of obstacle avoidance planning. The final success rate is approximately (37 successful times out of 44 grasping tasks). This solution performs poorly in a space with numerous, messy, narrow, and crowded obstacles, while in an environment with sufficient planning space, the AVO-QP solution has an excellent success rate of dynamic obstacle avoidance grasping.

[0069] On the other hand, as Figure 15 shown, the embodiments of the present invention provide a redundant robotic arm dynamic obstacle avoidance control device 3, which is connected to the redundant robotic arm 1. The device 3 includes a processor 30, a memory 32, and a computer program stored in the memory 32 and configured to be executed by the processor 30. When the processor 30 executes the computer program, it implements the redundant robotic arm dynamic obstacle avoidance control method as described in the above embodiments.

[0070] Exemplarily, the computer program may be divided into one or more modules / units, which are stored in the memory 32 and executed by the processor to implement the present invention. The one or more modules / units may be a series of computer program instruction segments capable of performing specific functions, and these instruction segments are used to describe the execution process of the computer program in the reticle repair control device. For example, the computer program may be divided into Figure 16 function modules in the reticle repair control device 1, wherein the data acquisition module 41, the constraint condition construction module 42, the cost function construction module 43, and the solution module 44 respectively execute the above steps S1 - S4.

[0071] The redundant robotic arm dynamic obstacle avoidance control device 3 may be a computing device such as a desktop computer, a notebook, a palm computer, or a cloud server. The redundant robotic arm dynamic obstacle avoidance control device 3 may include, but is not limited to, a processor 30 and a memory 32. Those skilled in the art can understand that the schematic diagram is only an example of the redundant robotic arm dynamic obstacle avoidance control device 3, and does not constitute a limitation on the redundant robotic arm dynamic obstacle avoidance control device 3. It may include more or fewer components than shown, or combine certain components, or different components. For example, the redundant robotic arm dynamic obstacle avoidance control device 3 may further include input / output devices, network access devices, a bus, etc.

[0072] The processor 30 may be a central processing unit (CPU), or may also be other general - purpose processors, digital signal processors (DSPs), application - specific integrated circuits (ASICs), field - programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general - purpose processor may be a microprocessor or the processor may also be any conventional processor, etc. The processor 30 is the control center of the redundant robotic arm dynamic obstacle avoidance control device 3, and connects various parts of the entire redundant robotic arm dynamic obstacle avoidance control device 3 through various interfaces and lines.

[0073] The memory 32 can be used to store the computer programs and / or modules. By running or executing the computer programs and / or modules stored in the memory 32, and invoking the data stored in the memory 32, the processor 30 realizes various functions of the redundant robotic arm dynamic obstacle avoidance control device 3. The memory 32 mainly includes a program storage area and a data storage area. Among them, the program storage area can store an operating system, application programs required for at least one function (such as a graphic recognition function, a graphic stacking function, etc.); the data storage area can store data created according to the use of the control device (such as graphic data, etc.). In addition, the memory 32 can include a high-speed random access memory, and can also include a non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, at least one magnetic disk storage device, a flash memory device, or other volatile solid-state storage devices.

[0074] If the functions described in the embodiments of the present invention are implemented in the form of software function modules or units and sold or used as independent products, they can be stored in a computer-readable storage medium that can be read by a computing device. Based on such an understanding, to implement all or part of the processes in the above-mentioned method embodiments, the present invention embodiments can also be completed by instructing relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium. When the computer program is executed by the processor 30, the steps of the above-mentioned method embodiments can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in the form of source code, object code, executable file, or some intermediate form, etc. The computer-readable medium can include: any entity or device capable of carrying the computer program code, a recording medium, a USB flash drive, a mobile hard disk, a magnetic disk, an optical disc, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium, etc. It should be noted that the content included in the computer-readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, the computer-readable medium does not include electrical carrier signals and telecommunication signals.

[0075] On the other hand, the embodiments of the present invention provide a computer-readable storage medium. The computer-readable storage medium includes a stored computer program. When the computer program runs, it controls the device where the computer-readable storage medium is located to execute the redundant robotic arm dynamic obstacle avoidance control method as described in the above embodiments.

[0076] The various embodiments in this specification are described in a progressive manner. Each embodiment focuses on the differences from other embodiments. For the same or similar parts among the various embodiments, reference can be made to each other.

[0077] The embodiments of the present invention have been described above in conjunction with the accompanying drawings. However, the present invention is not limited to the above specific embodiments. The above specific embodiments are merely illustrative and not restrictive. Under the inspiration of the present invention, those of ordinary skill in the art can also make many forms without departing from the spirit of the present invention and the scope protected by the claims. All of these fall within the protection scope of the present invention.

Claims

1. A redundant robotic arm dynamic obstacle avoidance control method, characterized in that: The method comprises the following steps: Acquire monitoring data of the redundant robotic arm in each control cycle collected and transmitted by sensors on the redundant robotic arm, wherein the monitoring data includes an initial joint angle of each joint of the redundant robotic arm and a closest point between the redundant robotic arm and an obstacle; Based on the initial joint angle and the closest point, the constraint conditions of the redundant manipulator in the current control cycle are constructed, wherein the constraint conditions include a target tracking constraint function based on the initial joint angle and about the joint speed, a dynamic obstacle avoidance constraint function based on the closest point and about the joint speed, a physical constraint function based on the speed escape method combined with the initial joint angle, and a self-collision constraint function based on the initial joint angle and used to prevent each joint of the redundant manipulator from colliding with itself, wherein the target tracking constraint function is an original tracking constraint function constructed based on the initial joint angle and is corrected by a pre-constructed feedback function and an avoidance vector; Constructing a cost function of the redundant manipulator with respect to the initial joint angle, combining the cost function with the constraint condition to obtain a set of quadratic programming equations of the redundant manipulator; and The RNN network model is used to solve the quadratic programming equations to obtain the target joint speed of each joint of the redundant robotic arm in the current control cycle.

2. The redundant manipulator dynamic obstacle avoidance control method according to claim 1, characterized in that: The feedback function is pre-constructed according to the following steps based on the difference between the initial pose and the expected pose of the end effector of the redundant manipulator in the current control cycle: Set the feedback function to , where X represents the current pose of the end effector of the redundant manipulator, represents the desired pose of the end effector of the redundant manipulator, and γ is the gain coefficient; The rotation matrix calculation based on the redundant manipulator , and convert the calculated error into a six-dimensional pose in Cartesian space, then the final feedback function is expressed as: ,in and Represent the current pose and expected pose of the end effector of the redundant manipulator, respectively. Represents the mapping from the homogeneous transformation matrix to the six-dimensional pose in Cartesian space.

3. The redundant manipulator dynamic obstacle avoidance control method according to claim 2, characterized in that: The avoidance vector is a vector that modifies the original tracking constraint function after the end effector enters the preset avoidance threshold range of the obstacle and is tangent to the vector between the obstacle and the end effector. The construction process of the avoidance vector specifically includes: Set avoidance vector , set the reference unit vector of the vector from the nearest point on the obstacle to the end effector to , the avoidance vector and the reference unit vector Tangent; Using the reference unit vector As the Z axis, a new three-dimensional coordinate system is constructed. The new three-dimensional coordinate system is expressed as: ,but The plane where it is located coincides with the XY plane of the three-dimensional coordinate system, The direction in the three-dimensional coordinate system is represented by the target unit vector , using trigonometric functions to optimize the target unit vector , then the target unit vector is expressed as ; Set the target unit vector and The rotation matrix for the transformation is R, where the rotation angle in the rotation matrix R is β and the rotation axis is and perpendicular to the target unit vector and , then the rotation angle β is expressed as: ； The rotation axis is It is expressed as: ； Based on the Rodriguez rotation formula, the rotation matrix R is expressed as: ； The target unit vector Transform to The transformation formula is expressed as: ； in, The size of the obstacle is related to the distance from the nearest point on the obstacle to the end effector. and the speed of the obstacle If relevant, It can be expressed as: ； in, For adjustment A positive constant of size; Set the unit vector from the end effector of the redundant manipulator to the target to be grasped as , based on the avoidance principle that is conducive to approaching the target to be grasped, a solution function is constructed: ； Among them, solving the The optimal α when the maximum value is obtained, and the avoidance vector is constructed based on the optimal α .

4. The redundant manipulator dynamic obstacle avoidance control method according to claim 3, characterized in that: The original tracking constraint function is expressed as: ,in, and Respectively and The first-order derivative with respect to time t represents the velocity of the redundant manipulator in Cartesian space and joint space, represents the Jacobian matrix, θ represents the initial joint angle; The target tracking constraint function is expressed as: ; The construction process of the dynamic obstacle avoidance constraint function specifically includes: Set the closest point on the redundant robot To the point on the obstacle The distance is , setting the distance threshold interval between the redundant manipulator and the obstacle [d2, d1], the obstacle avoidance constraint of the speed level of the redundant manipulator can be expressed as: , , where e is a scalar and , is a positive constant used to adjust the convergence speed of the obstacle avoidance constraint, ,in, Indicates the point on the redundant robot arm Point to the obstacle The unit vector of , Represents the point on the redundant robot arm speed; Based on the kinematic principle Represented as a point The Jacobian matrix And the product of the velocities of each joint of the redundant robot arm: ,in, ,and Only with redundant robot Related to previous joints; and make , , then the dynamic obstacle avoidance constraint function is expressed as: ; The physical constraint function is expressed as: ,in, and They represent the upper and lower limits of the motion angles of each joint of the redundant robotic arm, and Respectively represent the upper limit and lower limit of the angular velocity of each joint of the redundant manipulator; The construction process of the self-collision constraint function specifically includes: Set the controlled point on the redundant robot arm to The collision avoidance point is , then by controlling the controlled point speed to prevent it from colliding with the avoidance point The collision avoidance function when a collision occurs is expressed as: ,make , then the self-collision constraint function is expressed as .

5. The redundant manipulator dynamic obstacle avoidance control method according to claim 4, characterized in that: The cost function of constructing the redundant manipulator with respect to the initial joint angle specifically includes: Based on the initial joint angles of the current control cycle and the previous control cycle, the minimized velocity norm function and the smoothed joint velocity function of each joint of the redundant manipulator are respectively constructed, and the minimized velocity norm function is expressed as: ； The smooth joint velocity function is expressed as: ； Where T represents the control period, Respectively represent the joint speed of each joint of the redundant manipulator in the previous control cycle of the current control cycle; and The cost function is obtained by combining the minimized velocity norm function and the smooth joint velocity function, which is specifically expressed as: ； in, is a positive constant that adjusts the evaluation index.

6. The redundant manipulator dynamic obstacle avoidance control method according to claim 5, characterized in that: The quadratic programming equations of the redundant manipulator are expressed as: ； The dynamic obstacle avoidance constraint function and the self-collision constraint function are combined, and , .

7. The redundant manipulator dynamic obstacle avoidance control method according to claim 6, characterized in that: The use of the RNN network model to solve the quadratic programming equations to obtain the target joint speed of each joint of the redundant manipulator in the current control cycle specifically includes: The Lagrangian function is constructed based on the quadratic programming equations, which is specifically expressed as: ； Based on the KKT condition, the optimal solution of the Lagrangian function satisfies the following optimal solution relationship, which is specifically expressed as: ； in, represents the projection operator; Based on the optimal solution relational expression, calculation is performed to obtain the optimal solution equation group, which is specifically expressed as: ； in, To influence the learning rate of convergence speed, and They are the state variables that affect target tracking and obstacle avoidance respectively; Based on the principle of parallel computing, an RNN network model is constructed to solve the optimal solution equation group to obtain the target joint speed of each joint of the redundant robotic arm in the current control cycle.

8. The redundant manipulator dynamic obstacle avoidance control method according to claim 1 or 2, characterized in that: The monitoring data also includes a depth image of the target to be grasped by the redundant mechanical arm, and the method further includes: Selecting a target frame of the target to be captured from the depth image based on a preset target detection algorithm, and calculating and obtaining the spatial three-dimensional coordinates of the target to be captured based on the depth information of the depth image; Obtaining a current position and posture of the end effector of the redundant robotic arm in a current control cycle; and The desired posture of the end effector is solved and corrected based on the current posture, the target frame and the spatial three-dimensional coordinates of the target to be grasped.

9. A redundant mechanical arm dynamic obstacle avoidance control device, connected to the redundant mechanical arm, characterized in that: The device includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, and the processor implements the redundant robotic arm dynamic obstacle avoidance control method according to any one of claims 1 to 8 when executing the computer program.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored computer program, wherein when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute the redundant robotic arm dynamic obstacle avoidance control method according to any one of claims 1 to 8.

Images