Robot Action Program Generation Using Reusable Skills and Tasks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing robot programming systems lack efficiency in generating action programs due to limitations in reusing skills and tasks across different action reference coordinates and executing entities, leading to inflexible and time-consuming program creation processes.
Innovation Solution
A program generation device and method that generate skills, tasks, and masters by storing and associating relative actions with action reference coordinates, allowing for flexible reuse and combination to create action programs for robots, enabling efficient program generation by reusing skills and tasks across various action reference coordinates and executing entities.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If robot programming systems use fixed action reference coordinates for skills and tasks, then program execution is straightforward, but reusing skills and tasks across different coordinates becomes difficult
Solution Approach 1:
The patent implements a coordinate transformation mechanism that enables skills and tasks to function universally across different action reference coordinates. The master program includes coordinate transformation information that allows the same skill or task to be executed relative to multiple different coordinates without creating separate programs for each coordinate system.
Solution Approach 2:
The patent employs a nested structure where skills are contained within tasks, and tasks are contained within masters. This hierarchical nesting allows skills to be reused at multiple levels - a skill can be used in multiple tasks, and a task can be used in multiple masters, creating efficient reuse without increasing overall program complexity.
2Adaptability or versatility
If robot programming systems create separate programs for each executing entity, then program execution is clear and unambiguous, but reusing skills and tasks across different executing entities becomes time-consuming
Solution Approach 1:
The patent creates executing-entity-agnostic skills and tasks that can be assigned to multiple different executing entities. The master program contains the execution logic and coordinates entities rather than embedding entity-specific code in each skill or task, allowing the same skill-task combination to be reused across multiple robots or actuators without duplication.
3Measurement precision
If robot programming systems use absolute coordinates for all actions, then position precision is high, but flexibility in reusing actions at different positions is reduced
Solution Approach 1:
The patent segments the coordinate system into action reference coordinates (for precise positioning) and relative coordinates (for flexible reuse). Skills and tasks are defined with relative coordinate offsets from action reference points, allowing the same skill to be precisely positioned at different locations by simply changing the reference coordinate without modifying the skill's internal coordinate data.
Data Source
AI summary
A generation device for generating an action program for a robot based on an operation of a user includes processing circuitry that generates skill information including skills each corresponding to a relative robot action, stores the skill information in a skill database, generates a task including skills each associating with action reference coordinates that serve as a reference for the relative robot action, stores the task in a task database, generates a master including tasks associating with the robot, and stores the master in a master database.