Robot Calibration via Iterative Mechanical Error Identification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional robot calibration techniques face challenges in achieving absolute accuracy due to insufficient identification of mechanical error parameters, particularly when data is limited or not available for all necessary positions, leading to suboptimal tip position calculation.
Innovation Solution
A calibration system and method that includes a data storage unit, a parameter selecting unit, a parameter identifying unit, and an identification result evaluating unit, which iteratively select, identify, and evaluate mechanical error parameters to minimize the difference between actual and theoretical tip positions, repeating the process until a predetermined criterion is met.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If all mechanical error parameters are sequentially identified in groups (attitude parameters, linear motion direction parameters, and translational parameters), then the calibration process follows a systematic approach, but the absolute accuracy of the robot tip cannot be sufficiently improved when data is limited or unavailable for all necessary positions
Solution Approach 1:
The calibration system dynamically adapts the identification process by switching between different identification methods based on data availability. When measurement data is sufficient, the system uses identification with measurement data; when data is insufficient, it automatically transitions to identification without measurement data, ensuring the calibration can proceed under varying data conditions while maintaining accuracy improvement
Solution Approach 2:
The system changes the identification approach based on the availability of measurement data. It selectively applies different identification algorithms (with or without measurement data) and adjusts the optimization criteria accordingly, allowing the calibration to effectively utilize available data while compensating for missing data through alternative identification methods
2Productivity
If conventional calibration methods are used with limited measurement data, then the calibration process can be completed, but the identification of mechanical error parameters is insufficient leading to suboptimal tip position calculation
Solution Approach 1:
The calibration process is segmented into multiple stages: initial identification without measurement data to obtain baseline parameters, followed by refined identification with measurement data to improve accuracy. This segmentation allows the system to make progress even with limited data while systematically improving precision when data becomes available
Solution Approach 2:
The system performs preliminary identification of mechanical error parameters using available data before final calibration. This preliminary action establishes initial parameter values that serve as a foundation for subsequent refinement, ensuring that calibration can proceed productively even with incomplete measurement data
Data Source
AI summary
A calibration system of a robot includes a data storage unit that stores a number of pieces of data in which actual measured position information obtained by actually measuring a tip position of the robot and information indicating a state of the robot upon actual measurement are combined, a parameter selecting unit that selects mechanical error parameters to be identified from parameters indicating mechanical errors of the robot, a parameter identifying unit that identifies each of the mechanical error parameters using the stored data and information of the selected mechanical error parameters, and an identification result evaluating unit that evaluates an identification result, and, in the case where evaluation does not satisfy a predetermined criterion, selection of different mechanical error parameters by the parameter selecting unit, identification and evaluation are repeated.