A six-axis robot trajectory planning method based on B-spline

Through the six-axis robotic arm trajectory planning method based on B-spline, the problems of global speed control and interpolation speed error in spline curve planning are solved, and the smooth motion and efficient force control of the robotic arm are realized to meet the needs of complex processes.

CN116079714BActive Publication Date: 2024-11-12TUOJIXIN ANIMATION TECHNOLOGY (HUIZHOU) CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202211539346.X
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-12-02
Publication Date
2024-11-12
Estimated Expiration
2042-12-02

AI Technical Summary

Technical Problem

Spline curve planning is difficult to meet process requirements in terms of global speed control. When the long-distance curvature changes in a large range, the interpolation speed error is large, and the planning project is not convenient for adding force level hybrid control.

Method used

Using the six-axis robotic arm trajectory planning method based on B-spline, the global speed control and interpolation parameter calculation are realized by fitting the end motion trajectory, calculating the motion speed prospect, planning the trajectory interpolation and post-processing, combined with Simpson numerical integral and Taylor Expand and Adams implicit methods and other technologies to optimize speed planning and interpolation.

Benefits of technology

It realizes smoothness and low impact of the end movement of the robot arm, meets the process requirements such as spraying and grinding, reduces interpolation speed errors, and improves the real-time and calculation efficiency of force control.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN116079714B_ABST
    Figure CN116079714B_ABST
Patent Text Reader

Abstract

The invention discloses a six-axis robot trajectory planning method based on B-spline. Firstly, control point parameters and fitting methods are customized, and corresponding B-spline equations are fitted according to the control point parameters; by performing Simpson numerical integration on the B-spline parameter equation, the length of each piecewise function curve and the total length of the B-spline curve are obtained, the curvature of the B-spline parameter equation is sampled, the minimum running time is calculated according to the sampling speed, the interpolation time of the minimum integer is obtained by dividing the minimum interpolation cycle, and the maximum limit speed of each segment is inversely obtained; finally, according to the acceleration and the acceleration limit, two convolutions are performed to obtain smooth speed planning, and the interpolation parameters and interpolation points are obtained to complete the planned trajectory interpolation; finally, the planned interpolation points are post-processed, and points are centrally taken from the interpolation points to send actuators, and the modular process is convenient to replace with good adaptability, and the global linear speed forward planning is used to make the movement of the robot end smooth and the impact small.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of robot trajectory planning, and in particular to a six-axis robot arm trajectory planning method based on B-spline. Background Art

[0002] In the motion control of robotic arms, commonly used planning methods are divided into joint motion planning and end-position planning, among which the end-position motion planning commonly used includes linear motion, arc motion, circle motion and spline trajectory motion. Usually in the tracking of complex curves and surfaces in robotic arm spraying, grinding, massage and assembly, the current better method is to use the method of constructing B-splines. B-splines have the advantages of parameterization, geometric invariance, and local support. They are widely used in the industrial field. The curvature of high-order B-splines changes smoothly and has little impact on the system. However, the current spline curve planning has problems such as global speed control is difficult to meet the processability, the interpolation speed error is large when the curvature variation range is large over long distances, and the planning project is not convenient to add force-position mixed control. Summary of the invention

[0003] The present invention provides a six-axis robot arm trajectory planning method based on B-spline, which solves the problems of the current spline curve planning, such as the difficulty of global speed control in satisfying process performance, large interpolation speed error when the curvature variation range is large over long distances, and the inconvenience of adding force-position mixed control to the planning project.

[0004] To achieve the above object, the present invention provides the following technical solution: a six-axis robot arm trajectory planning method based on B-spline, comprising the following steps:

[0005] S1. First, fit the end motion trajectory, customize the control point parameters and fitting method, and fit the corresponding B-spline equation according to the control point parameters;

[0006] S2. By performing Simpson numerical integration on the B-spline parametric equation obtained in the previous step, the length of each piecewise function curve and the total length of the B-spline curve are obtained. At the same time, the curvature of the B-spline parametric equation is sampled at a given sampling interval to obtain the trajectory motion parameters, which provide content for subsequent speed forward planning and interpolation parameter calculation;

[0007] S3, plan the motion speed in advance, first calculate the minimum running time according to the sampling speed, then find the minimum integer through the minimum interpolation period, then find the maximum speed limit of each segment through the minimum integer interpolation time, and finally convolve twice according to the jerk and acceleration limit to obtain a smooth speed plan;

[0008] S4, according to the known B-spline curve equation and speed planning, and the interpolation parameters and interpolation points are calculated to complete the planned trajectory interpolation;

[0009] S5. Finally, the planned interpolation points are post-processed to obtain the Cartesian space interpolation pose, and points are selected from the interpolation point set and sent to the actuator.

[0010] Preferably, the control point parameters include position, posture and weight factor.

[0011] Preferably, the B-spline parameter equation C(u) is:

[0012]

[0013] where w i is n+1 weight factors corresponding to the control vertices, and the factor is 1 for quasi-uniform planning; d i There are n+1 control vertices; N i,p (u) is the node vector U = [u0,u1,u2,…,u n+p+1 The expression of the i-th p-th basis function on ] is:

[0014]

[0015] The final output is the B-spline parametric equation C(u).

[0016] Preferably, the B-spline parameter equation result includes the curvature radius of the position part and the curvature radius of the attitude part, wherein the position curvature radius is in meters and the attitude curvature radius is in radians.

[0017] Preferably, the Simpson numerical integration formula is:

[0018]

[0019] Where a and b are the upper and lower limits of the Simpson sampling interval parameters, and C(u) is the B-spline parameter equation.

[0020] Preferably, the convolution formula is:

[0021]

[0022] Among them, f(τ) in the first convolution is the speed limit obtained in the second part of the process, and the final speed limit is obtained after the interpolation interval specification; g(x-τ) is the acceleration signal:

[0023] g(x-τ)=1 / (Vmax / Acc / Ts)*ones(1,(Vmax / Acc / Ts))

[0024] In the above formula, Vmax is the maximum speed in the entire speed planning, Acc is the given acceleration, and Ts is the interpolation interval time;

[0025] In the second convolution, f(τ) is the planned speed obtained after the first convolution;

[0026] g(x-τ) is the acceleration signal:

[0027] g(x-τ)=1 / (Acc / Jerk / Ts)*ones(1,(Acc / Jerk / Ts))

[0028] In the above formula, Acc is the given acceleration, Jerk is the given jerk, and Ts is the interpolation interval time.

[0029] Preferably, in S4, the planned trajectory interpolation includes the following steps:

[0030] (1) When the interpolation period interval is small, the first-order Taylor expansion is used to find the parameters. The parameters are substituted into the equation to find the interpolation point, the first-order derivative and the second-order derivative of the interpolation point. Then, the speed error is calculated according to the bow height error formula and added to the speed of the interpolation point to get the actual speed. Then, the planned speed of the interpolation point is subtracted from the actual speed. When the speed difference is less than the specified error ε, the next step is performed. Otherwise, the parameter u is compensated iteratively by the ratio of planned speed / actual speed until the whole planning is completed.

[0031] (2) When the interpolation period interval is long, the first-order Taylor expansion is replaced by the n-order Taylor expansion, where n>1, and after calculating n+1 values, the Adams explicit and implicit method or the Runge-Kutta method is used to calculate the subsequent interpolation points;

[0032] (3) When the ratio of velocity to curvature radius is large, the binary iteration is used to sacrifice the number of iterations to improve stability;

[0033] (4) According to the speed V planned in S3, the relationship between the speed V and the parameter u about the parametric equation C(u) is established to obtain the interpolation parameter u, which is then brought into the B-spline parametric equation C(u) to obtain the interpolation posture, and it is iterated using the first-order Taylor expansion method;

[0034] (5) Interpolation parameter u based on the previous moment i And its relation to speed V(u i ), find the next interpolation parameter u i+1 :

[0035]

[0036] Where u is the interpolation parameter, i is the interpolation count, and t is the time. Generally speaking, t i+1 -t i That is, the fixed interpolation period Ts, The relationship between the velocity V and the B-spline parameter equation C(u) is introduced as follows:

[0037]

[0038] Among them, V(u i ) is the planned speed at the i-th interpolation, divided by the parameter equation C(u) at u i The rate of change of the second norm at the i interpolation time is calculated as follows: i )’s interpolation parameter u’s rate of change;

[0039] (6) By increasing the Taylor expansion order, the expansion truncation error is reduced, and the compensation ER is calculated by the curvature radius r of the current interpolation point and the actual speed v, and the compensation is applied to the actual speed to reduce the bow height error;

[0040] (7) Calculate C(u i ) Radius of curvature Bow height error Corrected actual speed after adding error compensation

[0041] (8) When the interpolation interval Ts is small, the interpolation parameter u i Directly corrected to Where V(u i ) is the planned speed of the current interpolation point, and finally the planned parameter u is substituted into the parameter equation C(u) to obtain the planned interpolation pose.

[0042] Preferably, in S5, the post-processing process includes, after a set of interpolation points of Cartesian space pose of B-spline planning has been obtained, starting to take points from the interpolation point set and sending them to the actuator, and choosing whether to enter the force control adjustment module, making secondary modifications to the interpolation points, and then inversely solving the Cartesian pose of the interpolation point into the joint angle θ according to the robot DH parameters, and finally sending them to each joint for execution.

[0043] Compared with the prior art, the present invention has the following beneficial effects:

[0044] 1. In the present invention, the global linear speed is forward-lookingly planned to make the end of the robot arm move smoothly with small impact, meeting the process requirements such as spraying and grinding.

[0045] 2. In the present invention, modular processes are used to facilitate replacement and good adaptability. Different process modules are replaced according to different computing and execution environments and requirements, such as selecting different patterns for fitting for online or offline programming.

[0046] 3. The present invention can plan a section of trajectory at one time and then process it through the force control algorithm, which greatly improves the real-time performance of force control and greatly shortens the control cycle. At the same time, in reciprocating working conditions, such as grinding and massaging, only one calculation is required, which can be used repeatedly to reduce the calculation load. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] The accompanying drawings are used to provide further understanding of the present invention and constitute a part of the specification. They are used to explain the present invention together with the embodiments of the present invention and do not constitute a limitation of the present invention.

[0048] In the attached picture:

[0049] Figure 1 It is the overall flow chart of the present invention;

[0050] Figure 2 It is the speed planning curve diagram of the present invention;

[0051] Figure 3 It is a schematic diagram of the NURBS curve bow height error compensation of the present invention. DETAILED DESCRIPTION

[0052] The preferred embodiments of the present invention are described below in conjunction with the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0053] Example: Figure 1 As shown in the figure, a trajectory planning method for a six-axis robot arm based on B-spline is shown. The overall process includes five parts: fitting the end motion trajectory, obtaining trajectory motion parameters, forward planning of motion speed, planning trajectory interpolation, and trajectory interpolation post-processing.

[0054] Part 1: Fitting the end motion trajectory, the process of fitting the corresponding B-spline equation according to the control point parameters, and the control points and fitting methods are customized. Figure 1 As shown in the overall process diagram, the control point position and posture are input, and the control points are densified by finding additional control points and making the curve pass through the control points, and then the control points are fitted using the 4th-order improved quasi-uniform B-spline to obtain the segmented B-spline parameter equation C(u); or when each control point and the node vector are known, the new control points passing through the original function curve are inversely calculated to obtain the parameter equation of the non-uniform B-spline curve; the method of only inputting the control point position and posture is suitable for robot online programming, which is conducive to less input and convenient operation; the method of inputting control points and node vectors is suitable for robot offline programming, which is conducive to richer control fitting information and more accurate description of the desired curve;

[0055] The control point parameters include its position (xyz), posture (rx ry rz) and weight factor ({w}).

[0056] B-spline parametric equation C(u):

[0057]

[0058] where wi is n+1 weight factors corresponding to the control vertices, and the factor is 1 for quasi-uniform planning; d i There are n+1 control vertices; N i,p (u) is the node vector U = [u0,u1,u2,…,u n+p+1 The expression of the i-th p-th basis function on ] is:

[0059]

[0060] The final output is the B-spline parametric equation C(u).

[0061] Part 2: The acquisition of spline trajectory parameters is to provide content for subsequent forward-looking speed planning and interpolation parameter calculation; on the one hand, the length of each piecewise function curve and the total length of the B-spline curve are obtained by performing Simpson numerical integration on the B-spline parameter equation obtained in the previous step; on the other hand, the curvature of the B-spline parameter equation is sampled at a given sampling interval. The results of the B-spline parameter equation include the curvature radius of the position part and the curvature radius of the attitude part. The position curvature radius is in meters and the attitude curvature radius is in radians. The attitude value is equivalent to the position curvature radius using the maximum acceleration of the robot motor and the connecting rod length as empirical coefficients for comparison. The minimum curvature radius is used as the forward-looking speed limit value at the same sampling point, and the Simpson adaptive integration method is used to calculate the line segment length, consuming less computing resources, thereby obtaining higher numerical accuracy; and by adopting the speed limit of the terminal posture, the requirements of the terminal linear speed in situations such as spraying, grinding, and massage are met, while reducing the impact on the robot system.

[0062] Simpson adaptive numerical integration formula:

[0063]

[0064] Where a and b are the upper and lower limits of the Simpson sampling interval parameters, and C(u) is the B-spline parameter equation.

[0065] The curvature velocity limit uses the method of equal centripetal acceleration value, and the formula is known:

[0066]

[0067] When the centripetal acceleration a is set to a constant value, the sampling curvature radius r determines the speed limit v, that is, the speed limit at each point of the trajectory curve is obtained. The value of a is set according to the performance of the robot arm and experience.

[0068] Part 3: Trajectory speed preview. Since the computer system is a discrete system with interpolation intervals, in order to accurately reach the position within the cycle, it is necessary to standardize the speed and time. First, calculate the minimum running time according to the sampling speed, then divide it by the minimum interpolation period to find the minimum integer, and then reverse the maximum speed limit of each segment through the minimum integer interpolation time. Finally, through the properties of convolution, the trajectory speed is quickly planned with very low computational cost and good adaptability. Acceleration and jerk are introduced, and the maximum speed limit of the step curve trajectory is convolved twice into a smooth speed planning curve, such as Figure 2 As shown;

[0069] Convolution formula:

[0070]

[0071] Among them, f(τ) in the first convolution is the speed limit obtained in the second part of the process, and the final speed limit is obtained after the interpolation interval specification; g(x-τ) is the acceleration signal:

[0072] g(x-τ)=1 / (Vmax / Acc / Ts)*ones(1,(Vmax / Acc / Ts))

[0073] In the above formula, Vmax is the maximum speed in the entire speed planning (m / s), and Acc is the given acceleration (m / s 2 ), Ts is the interpolation interval time (s).

[0074] In the second convolution, f(τ) is the planned speed obtained after the first convolution; g(x-τ) is the acceleration signal:

[0075] g(x-τ)=1 / (Acc / Jerk / Ts)*ones(1,(Acc / Jerk / Ts))

[0076] The above formula Acc is the given acceleration (m / s 2 ), Jerk is the given jerk (m / s 3 ), Ts is the interpolation interval time (s).

[0077] Part 4: Trajectory velocity interpolation, the process of finding interpolation parameters and interpolation points based on the known B-spline curve equation and velocity planning. According to the relationship between velocity and parameters on the curve equation, the Taylor expansion is used to quickly find the parameters. If the interpolation period interval is small, the first-order Taylor expansion can be used to find the parameters. The parameters are substituted into the equation to find the interpolation point, the first-order derivative and the second-order derivative of the interpolation point. Then, the velocity error is calculated according to the bow height error formula and filled into the interpolation point velocity to get the actual velocity. Then, the planned velocity of the interpolation point is subtracted from the actual velocity. When the velocity difference is less than the specified error ε, the next step is performed. Otherwise, the parameter u is compensated iteratively by the ratio of planned velocity / actual velocity until all planning is completed. When the interpolation period interval is long, the interpolation accuracy can be improved by improving the calculation accuracy and reducing the high-order error. For example, the first-order Taylor expansion is replaced by the n-order Taylor expansion, where n>1, and the subsequent interpolation points are calculated using the Adams explicit and implicit method or the Runge-Kutta method after calculating n+1 values, which not only improves the accuracy but also greatly reduces the calculation load. At the same time, when the ratio of velocity to curvature radius is large, the proportional correction is replaced by bisection iteration to improve the convergence performance and improve the stability by sacrificing the number of iterations.

[0078] According to the speed V planned in the third part, the relationship between it and the parameter u about the parametric equation C(u) is established to obtain the interpolation parameter u, and then bring it into the parametric equation C(u) to obtain the interpolation posture. Since it is difficult to obtain the analytical solution of the interpolation parameter u, especially under the high-order B-spline parametric equation; the first-order Taylor expansion method is used to iterate, based on the interpolation parameter u at the previous moment i And its relation to speed V(u i ), find the next interpolation parameter u i+1 :

[0079]

[0080] Where u is the interpolation parameter, i is the interpolation count, and t is the time. Generally speaking, t i+1 -t i That is, the fixed interpolation period Ts, The relationship between the velocity V and the B-spline parameter equation C(u) is introduced as follows:

[0081]

[0082] Among them, V(u i ) is the planned speed (m / s) at the i-th interpolation, divided by the parametric equation C(u) at u i The rate of change of the second norm at the i interpolation time can be calculated for the planned speed V(u i )’s interpolation parameter u’s rate of change.

[0083] like Figure 3As shown, since the interpolation interval Ts cannot be infinitely small, when the velocity V is larger and the curvature radius r is smaller, u i+1 The more the value of is, the less accurate it is. Its errors generally include iterative numerical error, expansion truncation error and bow height error. The expansion truncation error is caused by discarding high-order orders, which can be reduced by increasing the Taylor expansion order; the bow height error is also called contour error, which is caused by the fact that the interpolation points of the actual discrete system are linear motions rather than curves, and their length is less than the length of the fitting curve. The compensation ER can be calculated by the curvature radius r of the current interpolation point and the actual speed v, and the actual speed can be compensated to reduce the bow height error.

[0084] Calculate C(u i ) Radius of curvature Bow height error

[0085] Corrected actual speed after adding error compensation

[0086] When the interpolation interval Ts is small, the interpolation parameter u i Directly corrected to Where V(u i ) plans the speed for the current interpolation point.

[0087] Finally, the planned parameter u is substituted into the parametric equation C(u) to obtain the planned interpolation pose {xyz rx ry rz}.

[0088] Part 5: Post-processing of planned interpolation points is the process of post-processing the planned interpolation points. After obtaining a set of interpolation points {xyz rx ry rz} of Cartesian space poses planned by a B-spline, points are taken from the interpolation point set and sent to the actuator. You can choose whether to enter the force control adjustment module, such as constant force control, force exploration or admittance control, to make secondary modifications to the interpolation points, and then inversely solve the Cartesian pose of the interpolation point into the joint angle θ according to the robot DH parameters, and finally send it to each joint for execution.

[0089] Finally, it should be noted that the above description is only a preferred example of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments or replace some of the technical features therein by equivalents. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A six-axis robot arm trajectory planning method based on B-spline, characterized in that: The following steps are involved: S1. First, fit the end motion trajectory, customize the control point parameters and fitting method, and fit the corresponding B-spline equation according to the control point parameters; S2. By performing Simpson numerical integration on the B-spline parametric equation obtained in the previous step, the length of each piecewise function curve and the total length of the B-spline curve are obtained. At the same time, the curvature of the B-spline parametric equation is sampled at a given sampling interval to obtain the trajectory motion parameters, which provide content for subsequent speed forward planning and interpolation parameter calculation; S3, plan the motion speed in advance, first calculate the minimum running time according to the sampling speed, then find the minimum integer through the minimum interpolation period, then find the maximum speed limit of each segment through the minimum integer interpolation time, and finally convolve twice according to the jerk and acceleration limit to obtain a smooth speed plan; S4, according to the known B-spline curve equation and speed planning, and the interpolation parameters and interpolation points are calculated to complete the planned trajectory interpolation; S5. Finally, the planned interpolation points are post-processed to obtain the Cartesian space interpolation pose, and points are selected from the interpolation point set and sent to the actuator.

2. The six-axis robot arm trajectory planning method based on B-spline according to claim 1, characterized in that: The control point parameters include position, posture and weight.

3. The six-axis robot arm trajectory planning method based on B-spline according to claim 1, characterized in that: The B-spline parameter equation C(u) is: where w i is n+1 weight factors corresponding to the control vertices, and the factor is 1 for quasi-uniform planning; d i There are n+1 control vertices; N i,p (u) is the node vector U = [u0,u1,u2,…,u n+o+1 The expression of the i-th p-th basis function on ] is: The final output is the B-spline parametric equation C(u).

4. The six-axis robot arm trajectory planning method based on B-spline according to claim 1, characterized in that: The B-spline parameter equation result includes the curvature radius of the position part and the curvature radius of the attitude part, wherein the position curvature radius is in meters and the attitude curvature radius is in radians.

5. The six-axis robot arm trajectory planning method based on B-spline according to claim 1, characterized in that: The Simpson numerical integration formula is: Where a and b are the upper and lower limits of the Simpson sampling interval parameters, and C(u) is the B-spline parameter equation.

6. The six-axis robot arm trajectory planning method based on B-spline according to claim 1, characterized in that: The convolution formula is: Among them, f(τ) in the first convolution is the speed limit obtained in the second part of the process, and the final speed limit is obtained after the interpolation interval specification; g(x-τ) is the acceleration signal: g(x-τ)=1 / (Vmax / Acc / Ts)*ones(1,(Vmax / Acc / Ts)) In the above formula, Vmax is the maximum speed in the entire speed planning, Acc is the given acceleration, and Ts is the interpolation interval time; In the second convolution, f(τ) is the planned speed obtained after the first convolution; g(x-τ) is the acceleration signal: g(x-τ)=1 / (Acc / Jerk / Ts)*ones(1,(Acc / Jerk / Ts)) In the above formula, Acc is the given acceleration, Jerk is the given jerk, and Ts is the interpolation interval time.

7. The six-axis robot arm trajectory planning method based on B-spline according to claim 1, characterized in that: In S4, the planned trajectory interpolation includes the following steps: (1) When the interpolation period interval is small, the first-order Taylor expansion is used to find the parameters. The parameters are substituted into the equation to find the interpolation point, the first-order derivative and the second-order derivative of the interpolation point. Then, the speed error is calculated according to the bow height error formula and added to the speed of the interpolation point to get the actual speed. Then, the planned speed of the interpolation point is subtracted from the actual speed. When the speed difference is less than the specified error ε, the next step is performed. Otherwise, the parameter u is compensated iteratively by the ratio of planned speed / actual speed until the whole planning is completed. (2) When the interpolation period interval is long, the first-order Taylor expansion is replaced by the n-order Taylor expansion, where n>1, and after calculating n+1 values, the Adams explicit and implicit method or the Runge-Kutta method is used to calculate the subsequent interpolation points; (3) When the ratio of velocity to curvature radius is large, the binary iteration is used to sacrifice the number of iterations to improve stability; (4) According to the speed V planned in S3, the relationship between the speed V and the parameter u about the B-spline parameter equation C(u) is established to obtain the interpolation parameter u, which is then brought into the B-spline parameter equation C(u) to obtain the interpolation posture, and it is iterated using the first-order Taylor expansion method; (5) Interpolation parameter u based on the previous moment i And its relation to speed V(u i ), find the next interpolation parameter u i+1 : Where u is the interpolation parameter, i is the interpolation count, and t is the time. Generally speaking, t i+1 -t i That is, the fixed interpolation period Ts, The relationship between the velocity V and the B-spline parameter equation C(u) is introduced as follows: Among them, V(u i ) is the planned speed at the i-th interpolation, divided by the B-spline parameter equation C(u) at u i The rate of change of the second norm at the i interpolation time is calculated as follows: i )’s interpolation parameter u’s rate of change; (6) By increasing the Taylor expansion order, the expansion truncation error is reduced, and the compensation ER is calculated by the curvature radius r of the current interpolation point and the actual speed v, and the compensation is applied to the actual speed to reduce the bow height error; (7) Calculate C(u i ) Radius of curvature Bow height error Corrected actual speed after adding error compensation (8) When the interpolation interval Ts is small, the interpolation parameter u i Directly corrected to Where V(u i ) is the planned speed of the current interpolation point, and finally the planned parameter u is substituted into the B-spline parameter equation C(u) to obtain the planned interpolation pose.

8. The six-axis robot arm trajectory planning method based on B-spline according to claim 1, characterized in that: In S5, the post-processing process includes, after a set of interpolation points of Cartesian space pose of B-spline planning has been obtained, starting to take points from the interpolation point set and sending them to the actuator, and choosing whether to enter the force control adjustment module, making secondary modifications to the interpolation points, and then inversely solving the Cartesian pose of the interpolation point into the joint angle θ according to the robot DH parameters, and finally sending them to each joint for execution.