Robot Workpiece Installation Order From Reverse-Action Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing machine learning methods for robot actions of picking and installing workpieces do not consider the action of installing the workpiece, leading to inefficiencies and potential re-holding, which hinders the efficiency of installation work.
Innovation Solution
A machine learning method that learns a reverse-order action of removing workpieces in a predetermined installation state and determines an installation order based on this action, considering holding positions and constraints, to prevent re-holding and optimize installation efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If the robot learns only the picking action without considering installation action, then the picking action can be optimized, but the workpiece may be re-held during installation, reducing overall efficiency
Solution Approach 1:
The patent applies reverse-order learning by teaching the robot the removal action in reverse sequence to determine the optimal installation order. Instead of directly learning the installation action, the system learns how to remove workpieces from the container in reverse order, which automatically provides the optimal installation sequence. This inversion approach resolves the contradiction by ensuring that workpieces are picked and installed in an order that prevents re-holding, thereby maintaining picking optimization while eliminating installation inefficiencies.
Solution Approach 2:
The patent performs preliminary learning of the reverse-order removal action before actual installation operations. By pre-learning the removal sequences and using this knowledge to determine installation orders, the system prepares optimal installation paths in advance. This preliminary action ensures that when the robot performs actual installation, it already has the optimal sequence determined, preventing re-holding and improving installation efficiency without compromising picking optimization.
2Speed
If the robot picks workpieces from bulk stacking without considering installation order, then picking can be performed quickly, but installation may require re-holding, increasing total time
Solution Approach 1:
The patent uses reverse-order learning where the robot learns removal actions in reverse sequence. This inversion provides the optimal installation order that matches the bulk picking capability. By determining installation sequences through reverse removal learning, the system ensures that workpieces picked quickly from bulk stacking can be directly installed without re-holding, thus maintaining high picking speed while eliminating time loss from re-holding operations.
Solution Approach 2:
The patent incorporates feedback mechanisms where the learned reverse-order removal actions are used to optimize the installation sequence. The system continuously refines the installation order based on the feedback from reverse-order learning, ensuring that the picking speed advantage is maintained while preventing re-holding situations. This feedback loop resolves the contradiction by dynamically adjusting the installation sequence to match the bulk picking capability.
3Productivity
If the robot learns installation action directly, then installation efficiency can be improved, but the complexity of the learning system increases
Solution Approach 1:
The patent reduces learning system complexity by inverting the approach: instead of directly learning the complex installation action, the system learns the simpler reverse-order removal action. This inversion simplifies the learning process while still achieving optimal installation efficiency. The reverse-order learning provides the installation sequence without requiring the system to directly model complex installation constraints, thus resolving the contradiction between installation efficiency and system complexity.
Solution Approach 2:
The patent introduces reverse-order removal learning as an intermediary step between simple picking and complex installation learning. This intermediary approach allows the system to determine optimal installation sequences without directly learning complex installation actions. The reverse-order learning serves as a mediator that translates simple removal knowledge into optimal installation planning, reducing overall system complexity while maintaining installation efficiency.
Data Source
AI summary
A machine learning method for learning an action of a robot including a hand to pick out a workpiece from a container containing a plurality of the workpieces stacked in bulk and install the workpiece such that the workpiece is in a predetermined installation state includes learning a reverse-order action of removing, by the hand, the workpiece in the predetermined installation state after completion of installation, and learning an installation order of the workpiece based on a learning result of the reverse-order action of removing the workpiece.