Industrial Machine Cell Task Sharing Using Learned Robot States
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional industrial machine cell systems face challenges in optimizing task sharing among multiple robots, leading to potential task overloads and imperfect task execution due to unique robot states and conditions, especially when one robot stops, resulting in workpiece misses or reduced productivity.
Innovation Solution
A machine learning device that observes state variables of industrial machines, learns optimal task sharing through reinforcement learning or supervised learning methods, and adjusts task distribution dynamically to maintain production volume and maximize task performance across robots, using a neural network to compute rewards and update value functions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If task sharing is determined uniformly among robots, then load balance is improved, but task execution reliability deteriorates due to robot-specific capacity limitations
Solution Approach 1:
The patent implements dynamic task ratio adjustment based on real-time robot states. The system continuously monitors robot availability and dynamically recalculates task allocations using learned models, transitioning from static uniform distribution to adaptive dynamic allocation that responds to changing system conditions.
Solution Approach 2:
The system changes the parameter of task ratio allocation from fixed uniform values to variable values optimized by machine learning models. The learned task ratios are adjusted based on robot capabilities, current workload, and state variables, transforming the allocation parameter from constant to optimized variable.
2Reliability
If task ratio is determined in advance considering robot states, then task execution reliability is improved, but system adaptability deteriorates due to enormous combination complexity
Solution Approach 1:
The system implements self-service through autonomous machine learning models that automatically learn optimal task allocations and adjust to new conditions without human intervention. The models continuously improve their predictions based on accumulated data, enabling the system to serve itself in optimizing task distribution across varying scenarios.
Solution Approach 2:
The patent incorporates feedback mechanisms where task execution results and robot state information are fed back to the learning models. This feedback loop enables continuous refinement of task allocation strategies, allowing the system to adapt to new patterns and conditions while maintaining reliable execution through learned optimizations.
3Ease of operation
If tasks are redistributed after robot failure, then load balance is improved, but productivity deteriorates due to system reconfiguration time
Solution Approach 1:
The system performs preliminary learning and preparation by continuously training models on robot state data and potential failure scenarios. When a robot fails, the pre-trained models can immediately recalculate optimal task allocations without requiring time-consuming real-time analysis, enabling rapid redistribution that maintains both balance and productivity.
4Reliability
If dynamic task ratio control is implemented, then task execution reliability is improved, but device complexity increases due to learning system requirements
Solution Approach 1:
The patent replaces complex mechanical control systems with intelligent software-based machine learning models. Instead of using intricate hardware mechanisms to adjust task ratios, the system uses computational models that learn optimal allocations from data, substituting physical complexity with information processing capability.
Data Source
AI summary
A machine learning device, which performs a task using a plurality of industrial machines and learns task sharing for the plurality of industrial machines, includes a state variable observation unit which observes state variables of the plurality of industrial machines; and a learning unit which learns task sharing for the plurality of industrial machines, on the basis of the state variables observed by the state variable observation unit.