Robot singular configuration avoidance control method and device, storage medium and system

Abstract

The invention discloses a robot singular configuration avoidance control method and device, a storage medium and a system.The robot singular configuration avoidance control method comprises the following steps that S1, a robot kinematics model is established based on a Denavit-Hartenberg parameter method, and the current joint angle of a robot is obtained; s2, calculating based on the current joint angle of the robot to obtain a Jacobian matrix; and S3, adjusting the Jacobian matrix according to the singular avoidance control strategy, calculating to obtain a joint speed, and sending a joint speed instruction based on the joint speed so as to control the robot to move according to the joint speed instruction. According to the technical scheme, the torque tracking precision is improved, particularly, the tracking error is reduced to 0.20 in a singular region, and the error growth rate is reduced to 0.020 s <-1 >.

Description

Technical Field

[0001] This invention relates to the field of robot motion control technology, specifically to a robot singular configuration avoidance control method, device, storage medium, and system. Background Technology

[0002] With the rapid development of industrial automation, intelligent manufacturing, smart healthcare, and logistics, robots are increasingly being used in various tasks, especially collaborative robots, which have attracted much attention due to their ability to work in conjunction with humans. In practical applications, robots often need to complete high-precision tasks such as complex trajectory tracking, precision assembly, and surgical operations, which places extremely high demands on the stability, accuracy, and safety of motion control.

[0003] In robot motion control, the singular configuration problem is a common technical bottleneck. When a robot is in certain specific configurations, its Jacobian matrix becomes rank deficient, causing the robot to lose the ability to move in one or more directions, and the joint velocity increases sharply or even approaches infinity, which in turn leads to problems such as instability of the control system, increased trajectory tracking error, and aggravated motion jitter.

[0004] Currently, methods for avoiding singular configurations mainly include the standard pseudo-inverse method and the fixed-damped least squares method (DLS). However, these methods still have the following shortcomings: For example, the standard pseudo-inverse method cannot be effectively solved near singular points, leading to sudden changes in joint velocity and a sharp drop in tracking accuracy, which fails to meet the requirements for high stability control.

[0005] For example, while the fixed-damping least squares method can suppress sudden velocity changes, its damping coefficient is fixed, making it difficult to achieve a performance balance between singular and normal regions, often leading to the introduction of unnecessary tracking errors in non-singular regions.

[0006] Furthermore, existing methods largely rely on empirical parameter adjustments, lacking the ability to dynamically respond to the robot's real-time motion state and making it difficult to adapt to complex and ever-changing working environments. Moreover, in high-safety-requirement scenarios such as collaborative robots and surgical robots, the stability and smoothness of existing methods remain insufficient, making it difficult to guarantee the safety and reliability of human-robot interaction.

[0007] Therefore, an intelligent avoidance control method is needed that can dynamically adapt to the robot's motion state, effectively suppress sudden speed changes in strange regions, and maintain high tracking accuracy in normal regions, so as to improve the robot's overall motion performance and operational safety. Summary of the Invention

[0008] To address the shortcomings of existing technologies, the present invention aims to provide a robot singular configuration avoidance control method, device, storage medium, and system. The robot singular configuration avoidance control method constructs a singular configuration avoidance strategy based on the robot's real-time motion state. It automatically suppresses sudden velocity changes when approaching singular regions to ensure smooth motion; in normal regions, it prioritizes maintaining high-precision trajectory tracking. By combining the Jacobian matrix condition number to determine the degree of configuration singularity in real time, a corresponding nonlinear mapping function is established to achieve autonomous adjustment of control parameters. This method does not rely on prior experience for parameter tuning, possesses good environmental adaptability and motion robustness, and significantly improves the robot's operational stability and human-robot interaction safety under complex working conditions.

[0009] To achieve the above objectives, the present invention proposes the following technical solution: In a first aspect, the present invention proposes a robot singular configuration avoidance control method, which includes the following steps: Step S1: Establish a robot kinematic model based on the Denavit-Hartenberg parametric method to obtain the current joint angles of the robot; Step S2: Calculate the Jacobian matrix based on the robot's current joint angles; Step S3: Adjust the Jacobian matrix according to the singularity avoidance control strategy and calculate the joint velocity. Issue joint velocity commands based on the joint velocity to control the robot to move according to the joint velocity commands. The singularity avoidance control strategy includes the following steps: The system identifies whether the robot is in a singular configuration region. When the robot is not in a singular configuration region, the standard pseudo-inverse method is used to adjust the Jacobian matrix and calculate the joint velocities. When the robot is in a singular configuration region, the fixed-damped least squares method and the adaptive term method are used to adjust the Jacobian matrix and calculate the joint velocities.

[0010] In the technical solution described in this invention, compared with the prior art, the advantage of the singularity avoidance control strategy of this case actually lies in the adaptive damping mechanism. Therefore, compared with the numerical iteration method, this case can improve the robustness at the singular point on the one hand, and on the other hand, when the Jacobian matrix is ​​close to the singularity, the adaptive DLS automatically increases the damping coefficient to avoid joint velocity explosion or numerical instability. Near the singular point, the Jacobian matrix is ​​irreversible, and the Newton method may diverge or generate huge joint velocities, leading to control failure.

[0011] Furthermore, the technical solution described in this invention can significantly improve real-time performance and computational efficiency. Specifically, since the numerical iterative method requires multiple iterations to converge and may get stuck in local optima, the adaptive DLS obtains an approximate solution through one-step calculation, making it more suitable for real-time control.

[0012] Preferably, in step S2, the Jacobian matrix is ​​calculated using the following formula: Equation (1) in, v This represents the generalized velocity vector of the end effector. q Indicates joint velocity. J ( q ) represents the Jacobian matrix.

[0013] Preferably, in step S3, determining whether the robot is in a singular configuration region involves the following steps: when If so, the robot is in a singular configuration region; when If so, the robot is not in a singular configuration region; Where rank(J(q)) represents the rank of the joint Jacobian matrix, m represents the end degrees of freedom, n represents the number of joints, and min(m,n) represents finding the minimum value between m and n.

[0014] It needs to be explained that different regions of singularity are distinguished based on the rank deficiency of the Jacobian matrix. Specifically, angular velocity Jacobian rank deficiency indicates wrist singularity, linear velocity Jacobian rank deficiency indicates shoulder singularity, and a complete Jacobian rank deficiency indicates shoulder singularity. The three strategies are for before entering the singular region and after entering the singular region, respectively. The standard pseudo-inverse method fails after the robot enters the singular state, so the standard pseudo-inverse method is used for pre-singularity control to prevent the robot from entering the singular state. Once the robot enters the singular state, the least damped least squares method is used for control. For simple low-tracking environments, the fixed-damped least squares method is used. For complex environments with high tracking accuracy and acceptable stability loss costs, the adaptive damped least squares method is used.

[0015] Preferably, when using the standard pseudo-inverse method, the joint velocity is obtained by calculating the following formula: Equation (2); in, q This represents the joint velocity, i.e., the robot's joint coordinate vector. J Represent the Jacobian matrix and establish a linear relationship between joint velocity and end effector velocity. J T express J The transpose of the matrix, v This indicates the desired speed of the robot's end effector.

[0016] Preferably, when using the damped least squares method, the joint velocity is obtained by calculating the following formula: Equation (3); in,q This represents the joint velocity, i.e., the robot's joint coordinate vector. J Represents the Jacobian matrix, J T express J The transpose of the matrix, v Indicates the desired terminal velocity. I Represents the identity matrix. λ Indicates the damping coefficient. λ Take a fixed value of 0.1.

[0017] Preferably, when using the adaptive method, the joint velocity is obtained by calculating the following formula: Equation (4); in, q Indicates joint velocity. v Indicates the desired terminal velocity. J a It is obtained by calculation using equations (5) and (6); Equation (5); Equation (6); In equation (5), J Represents the Jacobian matrix, J T Denotes the transpose matrix of J. v Indicates the desired terminal velocity. I Represents the identity matrix; In equation (6), ω represents operability.

[0018] Preferably, in step S3, when the robot moves according to the joint speed command, a smooth trajectory from the initial configuration to the target configuration is generated using the joint space linear interpolation method.

[0019] In a second aspect, the present invention provides a robot control device, which includes a memory and a processor; The memory is used to store computer programs; The processor is used to implement the above-described robot singular configuration avoidance control method when executing the computer program.

[0020] Thirdly, the present invention proposes a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the above-described robot singular configuration avoidance control method.

[0021] Fourthly, the present invention proposes a robot control system applicable to six-degree-of-freedom robots, the robot control system comprising: End effector; And the aforementioned robot control device.

[0022] Compared with the prior art, the present invention has the following advantages and beneficial effects: The technical solution described in this invention improves trajectory tracking accuracy by using small damping (typically 1×e) in non-singular regions. -4 ~ 1×e -2 To maintain high accuracy, damping can be appropriately increased in singular regions to suppress error accumulation. For example, when using an adaptive term method, the trajectory tracking error can be reduced from 0.36 in the standard pseudo-inverse method to 0.20, and the error growth rate can be reduced from 0.036s. -1 Reduced to 0.020 s -1 .

[0023] Furthermore, the technical solution described in this invention, when employing the adaptive term method strategy, dynamically adjusts the damping coefficient through operability ω, resulting in a significant reduction in the maximum joint velocity from 0.1830 rad / s using the standard pseudo-inverse method strategy to 0.0211 rad / s, a decrease of 88.5%. The standard deviation of acceleration is reduced from 0.12 rad / s² using the standard pseudo-inverse method strategy to 0.03 rad / s², resulting in smoother motion.

[0024] Furthermore, the stability score of the fixed-damping least squares strategy in the technical solution described in this invention is 0.619, while that of the adaptive term strategy is 0.553. Although slightly lower, it still maintains high stability. The damping coefficient changes continuously and gradually with operability (although segmented, it transitions smoothly within the threshold range), reducing joint velocity jumps. Since the adaptive mechanism automatically enhances damping when approaching singularities, it is more targeted, and the average joint velocity reduction ratio in the singular region reaches 45.9%, which is better than the 27.5% of the fixed-damping least squares strategy. Based on the adaptive mechanism of real-time operability feedback, there is no need to pre-calibrate the singular point position, and it can adapt to complex trajectories containing multiple singular points. Attached Figure Description

[0025] Figure 1 Visualization of different configurations of the PUMA560 robot, where A is the normal configuration, B is the wrist singularity, C is the elbow singularity, and D is the shoulder singularity; Figure 2 For the smoothness analysis of the control strategy; Figure 3 For regional performance comparison analysis; Figure 4 Analysis of adaptive damping strategy. Detailed Implementation

[0026] The present invention will now be described in detail with reference to specific embodiments. These embodiments will help those skilled in the art to further understand the present invention, but do not limit the invention in any way. It should be noted that those skilled in the art can make several adjustments and improvements without departing from the concept of the present invention. These all fall within the scope of protection of the present invention.

[0027] Example 1 In this embodiment, the PUMA560 industrial robot is used as the research object to systematically study the singular configuration problem encountered in its motion control.

[0028] The steps of the robot singular configuration avoidance control method adopted are as follows: Step S1: Establish a robot kinematic model based on the Denavit-Hartenberg parametric method to obtain the current joint angles of the robot; Step S2: Calculate the Jacobian matrix based on the robot's current joint angles; Step S3: Adjust the Jacobian matrix according to the singularity avoidance control strategy and calculate the joint velocity. Issue joint velocity commands based on the joint velocity to control the robot to move according to the joint velocity commands.

[0029] In the above scheme, the Jacobian matrix is ​​the core mathematical tool connecting the velocity mapping relationship between the robot's joint space and manipulator space. For a robot system with n degrees of freedom, the Jacobian matrix has a dimension of 6×n, where the first three dimensions correspond to the linear velocity mapping of the end effector, and the last three dimensions correspond to the angular velocity mapping. This structural characteristic allows the Jacobian matrix to completely describe the differential relationship between the robot's end effector's motion capability and joint motion. For an n-DOF robot, based on the DH parameters of the PUMA560 robot, the geometric Jacobian matrix can be constructed using the differential transformation method.

[0030] The Jacobian matrix is ​​shown below: Equation (1) in, The generalized velocity vector of the end effector. ,in, J v Indicates linear velocity Jacobian. J ω Jacobian represents angular velocity. j v This represents the linear velocity component of the end effector. j ω This represents the angular velocity component of the end effector.

[0031] Based on kinematic properties, the motion of a robot can be naturally decomposed into two independent components: translation and rotation. Therefore, the Jacobian matrix can be analyzed by dividing it into two parts: linear velocity and angular velocity.

[0032] Accordingly, in the technical solution described in this invention, the linear velocity Jacobian matrix can be written as follows: Equation (7) The angular velocity Jacobian matrix can be written as follows: Equation (8) in, x , y , z Represents the three-dimensional coordinates of the robot's coordinate system; , , This represents the partial derivative in the corresponding three-dimensional coordinate system.

[0033] A singular configuration is a robot configuration with a rank-deficient Jacobian matrix, in which the robot loses the ability to move in one or more directions. The mathematical criteria for singularity are as follows: Equation (9).

[0034] For a 6-DOF robot, when The robot is in a strange configuration.

[0035] Based on the structural characteristics of PUMA560, common singular configurations can be divided into three categories: wrist singularity, elbow singularity, and shoulder singularity.

[0036] Figure 1 Visualization of different robot configurations, where A is the normal configuration, B is the wrist singularity, C is the elbow singularity, and D is the shoulder singularity.

[0037] like Figure 1 As shown, A is the normal configuration, with joint angle parameters q1=0.500 rad, q2=0.785 rad, q3=1.047 rad, q4=0.200 rad, q5=0.785 rad, q6=0.100 rad, performance index condition number Inf, and operability 1.91e. -10 ; B is a wrist singularity with joint angle parameters q1=0.500 rad, q2=0.785 rad, q3=1.047 rad, q4=0.200 rad, q5=0.001 rad, q6=0.100 rad, performance index condition number Inf, and operability 6.90 e. -11 In other words, the mathematical condition is the angle of the 5th joint. The Jacobian property is that the angular velocity Jacobian matrix is ​​rank deficient. Physically, this means that the rotation axes of the three joints in the wrist are coplanar, resulting in the loss of one rotational degree of freedom. C represents the elbow singularity, with joint angle parameters q1=0.500 rad, q2=0.100 rad, q3=3.100 rad, q4=0.200 rad, q5=0.785 rad, q6=0.100 rad, performance index condition number Inf, and operability 6.17e. -10 In other words, the mathematical condition is that the third joint is close to fully extended or retracted, and the Jacobian property is that the linear velocity Jacobian matrix is ​​rank deficient. In physical terms, this means that the robotic arm is fully extended or folded, losing its ability to move in a certain direction. D represents a shoulder singularity, with joint angle parameters q1=0.000 rad, q2=0.100 rad, q3=0.100 rad, q4=0.200 rad, q5=0.785 rad, q6=0.100 rad, performance index condition number Inf, and operability 4.54 e. -10 In other words, the mathematical condition is that the first three joints are specially configured, and the Jacobian property is that the complete Jacobian matrix is ​​rank-deficient. The physical meaning is that the robot is at the boundary of the workspace, and the condition number of the Jacobian matrix increases sharply.

[0038] exist Figure 1 In this context, the condition number of the Jacobian matrix is ​​used to measure numerical stability, and it can be calculated using the following formula: Equation (9) in, and These are the maximum and minimum singular values ​​of the Jacobian matrix, respectively.

[0039] The operability ω is an indicator of a robot's flexibility, which can be calculated using the following formula: Equation (10) in, These are the singular values ​​of the Jacobian matrix. A higher operability indicates better maneuverability of the robot in that configuration. Near the singular configuration, the minimum singular value approaches zero, while the condition number approaches infinity.

[0040] When the robot approaches or is in a singular configuration, the standard inverse method fails. Therefore, in order to achieve stable and controllable motion near the singular region, the technical solution described in this invention adopts different singularity avoidance control strategies, namely the standard pseudo-inverse method, the fixed damped least squares method, and the adaptive term method.

[0041] When the robot is in a non-singular region, a standard pseudo-inverse singularity avoidance control strategy is adopted, that is, the Moore-Penrose pseudo-inverse of the Jacobian matrix is ​​used to map the end-effector space velocity to the joint space velocity, and the joint velocity is solved by the Moore-Penrose pseudo-inverse of the Jacobian matrix.

[0042] The formula used is as follows: Equation (2) However, when the robot fully enters the singular configuration (i.e., when the robot is in a singular region), Since it is a singular matrix, its pseudo-inverse cannot be calculated, and the control system fails.

[0043] When using the fixed-damped least squares method, the formula is as follows: Equation (3); It should be noted that the introduction of a damping coefficient to improve numerical stability in singular regions is essentially a trade-off between exact solutions and numerical stability in the fixed-damped least squares method. Adding a damping term ensures the matrix remains invertible, allowing for finite joint velocity solutions even in singular configurations. However, the performance of this algorithm is highly dependent on the choice of the damping coefficient, requiring a balance between velocity suppression and tracking accuracy.

[0044] To overcome the limitations of the fixed-damping least squares method, the technical solution of this invention incorporates an adaptive term (i.e., the adaptive term method). This dynamically adjusts the damping coefficient based on the operability metric, thereby allowing the damping coefficient to dynamically adjust according to the robot's real-time motion performance state, thus achieving intelligent performance balance. The joint velocity is obtained by calculating the following formula: Equation (4); in, q Indicates joint velocity. v Indicates the desired velocity of the joint. J a It is obtained by calculation using equations (5) and (6); Equation (5); Equation (6); In equation (5), J Represents the Jacobian matrix, J T express J transpose, v Indicates the end, I Represents the identity matrix ； In equation (6), ω represents operability.

[0045] Figure 4 Analysis of adaptive damping strategy.

[0046] like Figure 4 As shown, the design of the adaptive term is based on the strong correlation between operability and singular proximity. Operability A smaller value indicates that the robot is closer to a singular configuration, requiring a larger damping coefficient to suppress joint velocities. This can be achieved by setting an operability threshold. and This allows for segmented adjustment of the damping coefficient. It should be noted that the core advantage of the adaptive damping least squares strategy lies in its environmental adaptability. It achieves high tracking accuracy by using a small damping coefficient in the normal region, balances performance by using a medium damping coefficient in the transition region, and effectively suppresses velocity peaks by using a large damping coefficient in the singular region.

[0047] Furthermore, in step S4, a smooth trajectory from the initial configuration to the target configuration is generated using joint space linear interpolation. This method ensures the continuity of joint position and velocity, avoiding interference with the performance evaluation of the control strategy due to the inherent unevenness of the trajectory.

[0048] Example 2 In this embodiment, the MATLAB R2024b simulation platform is used. A PUMA560 robot model is built in the Robotics System Toolbox for kinematic and dynamic simulation. This includes core functions such as forward / inverse kinematics, Jacobian matrix calculation, and trajectory generation. The PUMA560 robot uses standard DH parameters to build its kinematic model. The simulation is performed using the `loadrobot("puma560")` function from the toolbox. The simulation time is t=10s, and the time step is [missing information]. The time vector is defined as .

[0049] The homogeneous transformation matrix is ​​as follows Equation (11) in, Indicates joint angle, This indicates the torsion angle of the connecting rod.

[0050] In this embodiment, to comprehensively evaluate the performance of the singularity avoidance strategy, a test trajectory containing singularities was designed. The results are shown in [reference needed]. Figure 2 and Figure 3 As shown. Figure 2 For the smoothness analysis of the control strategy. Figure 3 For regional performance comparison analysis; It should be noted that, Figure 2The curve containing the standard pseudo-inverse indicates that the singular avoidance control strategy adopted by this technical solution is the standard pseudo-inverse method, while the curve containing the fixed DLS indicates that the singular avoidance control strategy adopted by this technical solution is the fixed damped least squares method, and the curve containing the adaptive DLS indicates that the singular avoidance control strategy adopted by this technical solution is the adaptive term method.

[0051] Figure 3 In the control strategy diagram, bar 1 indicates that the singularity avoidance control strategy adopted by the technical solution is the standard pseudo-inverse method, while bar 2 indicates that the singularity avoidance control strategy adopted by the technical solution is the fixed damped least squares method, and bar 3 indicates that the singularity avoidance control strategy adopted by the technical solution is the adaptive term method.

[0052] Figure 2 and Figure 3 The joint acceleration norm and the mean joint velocity norm can be found in the formulas. as well as The calculations were performed using conventional techniques known to those skilled in the art, and therefore will not be elaborated upon further.

[0053] like Figure 2 and Figure 3 As shown, trajectory planning employs a joint-space linear interpolation method. The three control strategies exhibit significant differences in trajectory tracking accuracy. The standard pseudo-inverse method demonstrates good tracking performance in non-singular regions, but near singular points, the Jacobian matrix becomes ill-conditioned, leading to a sharp increase in solution error. Ultimately, the accumulated trajectory tracking error reaches 0.36 within a 10-second simulation. This error accumulation primarily stems from the sharp increase in the condition number of the Jacobian matrix near singular points, making numerical computation unstable. The fixed DLS method, by introducing a constant damping coefficient, effectively improves numerical stability near singular points, reducing the tracking error to 0.27. However, in non-singular regions, the presence of the damping coefficient introduces unnecessary conservatism, resulting in some loss of tracking accuracy. This accuracy loss may become a limiting factor in applications requiring high-precision tracking. The adaptive DLS strategy exhibits the best tracking performance, with a final error of only 0.20. This is thanks to its intelligent damping adjustment mechanism, which dynamically adjusts the damping coefficient based on operability indicators. In non-singular regions, it uses lower damping to maintain high-precision tracking, while in singular regions, it appropriately increases damping to ensure stability. Error growth rate analysis further validates the advantages of this strategy; the adaptive DLS's error growth rate is only 0.020s. -1 This is lower than the 0.036s of the standard pseudo-inverse method. -1 And 0.027s for fixed DLS -1 This superior error suppression capability makes the adaptive DLS strategy particularly suitable for industrial applications that require long-term stable operation.

[0054] Motion smoothness is another important indicator for evaluating the performance of control strategies. The standard pseudo-inverse method exhibits significant abrupt changes in joint velocity when traversing singular regions, with a maximum velocity norm reaching 0.1830 rad / s, while also experiencing severe acceleration fluctuations, peaking at 0.5 rad / s. 2 This indicates a significant discontinuity in its control output. This discontinuity not only affects trajectory tracking accuracy but may also cause additional wear on the mechanical structure. The fixed DLS method effectively suppressed velocity abrupt changes through the damping term, reducing the maximum joint velocity by 78.1% and the acceleration standard deviation by 0.05 rad / s². 2 Although it represents a significant improvement over the standard pseudo-inverse method, a certain degree of jitter still exists during motion, mainly because the fixed damping coefficient cannot adapt to the needs of different motion states. Adaptive DLS performs best in terms of motion smoothness, further reducing the maximum joint velocity to 0.0211 rad / s (a reduction of 88.5%), and the standard deviation of acceleration to only 0.03 rad / s². 2 This superior performance stems from its smooth damping transition mechanism, which makes joint velocity changes more continuous and natural, avoiding the jitter that may occur during mode switching in a fixed DLS strategy. Especially in the transition region, adaptive DLS ensures a smooth transition of the motion process through gradual damping adjustment.

[0055] In terms of stability, the fixed DLS strategy achieved a stability score of 0.619, slightly higher than the adaptive DLS's 0.553. This result indicates that although adaptive DLS leads in most performance metrics, its parameter switching mechanism introduces small fluctuations in the transition region, affecting stability. The standard pseudo-inverse method has the lowest stability score due to its severe fluctuations near singularities. Regarding real-time performance, the three strategies exhibit different computational complexity characteristics. The standard pseudo-inverse method has the lowest computational cost but requires an additional singularity detection mechanism. Fixed DLS has a slightly increased computational complexity, but it remains within acceptable limits for real-time control. Adaptive DLS has the highest computational cost due to the need for real-time calculation of operability and adjustment of damping coefficients, but it is still within acceptable limits for modern processors. In practical applications, a trade-off between performance and computational complexity needs to be struck based on specific requirements.

[0056] The specific embodiments of the present invention have been described above. It should be understood that the present invention is not limited to the specific embodiments described above, and those skilled in the art can make various modifications or variations within the scope of the claims, which do not affect the essence of the present invention.

Claims

1. A method for controlling singular configuration avoidance in a robot, characterized in that, The robot singular configuration avoidance control method includes the following steps: Step S1: Establish a robot kinematic model based on the Denavit-Hartenberg parametric method to obtain the current joint angles of the robot; Step S2: Calculate the Jacobian matrix based on the robot's current joint angles; Step S3: Adjust the Jacobian matrix according to the singularity avoidance control strategy and calculate the joint velocity. Issue joint velocity commands based on the joint velocity to control the robot to move according to the joint velocity commands. The singularity avoidance control strategy includes the following steps: The system identifies whether the robot is in a singular configuration region. When the robot is not in a singular configuration region, the standard pseudo-inverse method is used to adjust the Jacobian matrix and calculate the joint velocities. When the robot is in a singular configuration region, the fixed-damped least squares method and the adaptive term method are used to adjust the Jacobian matrix and calculate the joint velocities.

2. The robot singular configuration avoidance control method according to claim 1, characterized in that, In step S2, the Jacobian matrix is ​​calculated using the following formula: Equation (1) in, v This represents the generalized velocity vector of the end effector. q Indicates joint velocity. J ( q ) represents the Jacobian matrix.

3. The robot singular configuration avoidance control method according to claim 1, characterized in that, In step S3, it is determined whether the robot is in a singular configuration region. The specific operation is as follows: when If so, the robot is in a singular configuration region; when If so, the robot is not in a singular configuration region; Where rank(J(q)) represents the rank of the Jacobian matrix, and min(m,n) represents taking the smaller of m and n.

4. The robot singular configuration avoidance control method according to any one of claims 1 to 3, characterized in that, When using the standard pseudo-inverse method, the joint velocity is obtained by calculating the following formula: Equation (2); in, q Indicates joint velocity. J Represents the Jacobian matrix, J T To represent the transpose of the Jacobian matrix, v This indicates the desired velocity at the joint end.

5. The robot singular configuration avoidance control method according to any one of claims 1 to 3, characterized in that, When using the fixed-damped least squares method, the joint velocity is obtained by calculating the following formula: Equation (3); in, q Indicates joint velocity. J Represents the Jacobian matrix, J T express J transpose, v I represents the desired velocity at the joint end, and I represents the identity matrix. λ Indicates the damping coefficient. λ Take a fixed value of 0.

1.

6. The robot singular configuration avoidance control method according to any one of claims 1 to 3, characterized in that, When using the adaptive method, the joint velocity is obtained by calculating the following formula: Equation (4); in, q Indicates joint velocity. v Indicates the desired terminal velocity. J a It is obtained by calculation using equations (5) and (6); Equation (5); Equation (6); In equation (5), J Represents the Jacobian matrix, J T To represent the transpose of the Jacobian matrix, v Indicates the desired terminal velocity. I Represents the identity matrix ； In equation (6), ω represents operability.

7. The robot singular configuration avoidance control method according to any one of claims 1 to 3, characterized in that, In step S3, when the robot moves according to the joint velocity command, a smooth trajectory from the initial configuration to the target configuration is generated using the joint space linear interpolation method.

8. A robot control device, characterized in that, The robot control device includes a memory and a processor; The memory is used to store computer programs; The processor, when executing the computer program, implements the robot singular configuration avoidance control method as described in any one of claims 1 to 7.

9. A computer-readable storage medium, characterized in that, The storage medium stores a computer program, which, when executed by a processor, implements the robot singular configuration avoidance control method as described in any one of claims 1 to 7.

10. A robot control system, characterized in that, The robot control system is applicable to six-degree-of-freedom robots, and the robot control system includes: End effector; And the robot control device as described in claim 8.

Images