Robot Teaching by Extracting Essential Motions From Skilled Tasks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing robot teaching devices reproduce unnecessary operator motions, such as hand trembling and hesitation, reducing working efficiency and making it difficult to accurately replicate skilled tasks requiring precision.
Innovation Solution
A robot teaching device that acquires motion data, determines priority periods based on feature amounts indicating the operator's influence on task objects, and generates teaching data to reproduce only essential motions while omitting unnecessary ones.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the robot reproduces the operator's motion as it is, then the operator's intention is reproduced, but unnecessary motions such as hand trembling and hesitation are also reproduced, reducing working efficiency
Solution Approach 1:
The patent extracts only the essential motion components from the operator's original motion data. By analyzing motion trajectories and identifying key positional changes, the system separates necessary motions (that contribute to task completion) from unnecessary motions (such as hand trembling and hesitation). The robot then reproduces only the extracted essential motions, thereby maintaining accuracy of operator's intention while eliminating inefficiencies.
2Productivity
If the robot estimates and replans the motion to improve working efficiency, then unnecessary motions are eliminated, but it becomes difficult to completely reproduce the intention of the operator especially in skilled work requiring precision
Solution Approach 1:
The patent employs feedback mechanisms where the system continuously monitors and analyzes the operator's motion data, comparing original motions with reproduced motions. By using motion trajectory analysis and key point detection, the system provides feedback on which motion components are essential and which can be optimized. This feedback loop enables the robot to learn from the operator's skilled motions while maintaining precision, resolving the contradiction between efficiency improvement and accurate intention reproduction.
Data Source
AI summary
A robot teaching device and a robot teaching method with which a robot can be caused to carry out a series of tasks with high work efficiency. A feature quantity acquisition unit acquires feature quantity data indicating an amount of effect that an action of a worker has on work targets; an important point period determination unit determines, from periods for which action data is acquired and on the basis of the feature quantity data, periods in which the worker is performing actions essential to the series of tasks, as important point periods; and a teaching data generation unit generates, on the basis of the action data, teaching data to be input into robots so that the actions of the worker in the important point periods are reproduced and the actions of the worker outside the important point periods are not reproduced.