Robot Joint Control Using Deflection Error Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional robot control methods face challenges in achieving precise position and orientation accuracy due to trajectory errors, particularly in multi-joint robots with low stiffness, leading to significant correction errors that require convergent calculations within tight time constraints.
Innovation Solution
A robot controlling method that calculates error components in the driving direction and residual errors for each joint, using these to derive correction quantities for joint instruction values, reducing the need for convergent calculations and minimizing correction errors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If inverse mechanism calculation is used to calculate correction quantity of joint instruction values, then position and orientation error of robot end can be corrected, but calculation error increases due to simulation assumptions that ignore joint deflection
Solution Approach 1:
The patent segments the correction calculation into two distinct parts: (1) correction quantity calculation based on robot end position and orientation errors, and (2) deflection error calculation based on actuator driving forces and joint stiffness. This segmentation allows each part to be optimized independently, resolving the contradiction between computational simplicity and correction accuracy.
Solution Approach 2:
The patent introduces deflection error as an intermediary parameter that bridges the gap between actuator driving forces and joint position errors. By calculating deflection error based on joint stiffness and actuator forces, the system accurately captures the relationship between driving forces and position deviations without relying on inaccurate inverse mechanism simulations.
2Manufacturing precision
If convergent calculation is performed to reduce correction error, then correction accuracy improves, but calculation time increases beyond the 2ms constraint
Solution Approach 1:
The patent performs preliminary calculation of deflection error based on actuator driving forces and joint stiffness characteristics before the convergent correction calculation. By pre-calculating the deflection component, the subsequent convergent calculation starts from a more accurate initial state, reducing the number of iterations needed and ensuring the 2ms calculation constraint is met while maintaining high correction accuracy.
3Productivity
If restrictions of convergence determination are eased to reduce calculation time, then calculation speed increases, but correction error increases
Solution Approach 1:
By pre-calculating deflection error before the convergent loop, the patent provides a more accurate initial correction quantity that is closer to the final solution. This allows the convergence determination restrictions to be eased (reducing calculation time) while maintaining correction accuracy, because the preliminary deflection calculation has already captured a significant portion of the correction needed.
Solution Approach 2:
The patent changes the parameters used in correction calculation by explicitly incorporating joint stiffness and actuator driving forces to calculate deflection error. This parameter change transforms the correction approach from pure inverse mechanism simulation to a hybrid method that accounts for physical deflection, improving convergence behavior and allowing relaxed determination restrictions without sacrificing accuracy.
Data Source
AI summary
A controlling unit obtains an error in position and orientation of each joint of a robot. The controlling unit uses an error component in a driving direction of an actuator included in the error in position and orientation ui of the joint to obtain a first correction quantity, to obtain a residual error excluding the error component in the driving direction of the actuator from the error in position and orientation of the joint, and to obtain—an error in position and orientation of the end point of the robot based on the residual error of each joint. The controlling unit uses the error in position and orientation of the joint based on the error in position and orientation of the end point of the robot to obtain a second correction quantity Δqi, and uses the first correction quantity and the second correction quantity to correct a joint instruction value.