Machine Learning Device for Laser Robot Cycle Time Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
It is challenging to efficiently teach laser processing robots to achieve the shortest cycle time, especially when multiple robots are involved, as conventional methods are complex and difficult to optimize simultaneously.
Innovation Solution
A machine learning device that observes plasma light and processing sound from laser processing robots, calculates rewards based on cycle time and processing conditions, and updates a value function to learn and decide optimal processing orders, utilizing reinforcement learning and neural networks for efficient operation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional manual teaching methods are used to set processing orders and conditions, then operators can control robot operations, but it becomes extremely difficult to achieve optimal cycle times and the teaching process becomes overly complicated especially with multiple robots
Solution Approach 1:
The system enables self-learning through reinforcement learning where the machine learning device automatically optimizes processing orders and conditions without requiring manual teaching. The device learns optimal solutions by receiving rewards based on cycle time performance, eliminating the need for operators to manually program complex multi-robot sequences while achieving optimal productivity.
Solution Approach 2:
The system implements feedback mechanisms where the machine learning device receives reward signals based on actual cycle time measurements. This feedback loop allows the system to continuously learn and improve processing orders by comparing actual performance against targets, automatically adjusting to achieve optimal cycle times without manual intervention.
2Productivity
If multiple laser processing robots are used to increase processing capacity, then more welding can be performed simultaneously, but the teaching complexity increases dramatically as operators must coordinate which robots perform which operations in what order
Solution Approach 1:
The machine learning device automatically learns optimal task allocation and coordination for multiple robots through self-service learning. By observing state variables and receiving feedback on overall system cycle time, the device independently determines which robots perform which operations and in what sequence, eliminating the need for operators to manually coordinate complex multi-robot interactions.
Solution Approach 2:
The system changes the approach from manual parameter setting to automated learning by introducing reinforcement learning parameters. The machine learning device learns optimal processing orders by adjusting internal parameters based on reward signals, enabling efficient coordination of multiple robots without requiring operators to manually set complex coordination parameters.
3Productivity
If operators manually teach welding parts and welding order to achieve shortest cycle time, then some level of optimization can be achieved, but it is fundamentally difficult to reach the true optimal solution due to the complexity of the search space
Solution Approach 1:
The system uses reinforcement learning with feedback loops where the machine learning device receives reward signals based on cycle time performance. This allows the device to systematically explore the solution space and learn optimal processing orders through iterative improvement, overcoming the limitations of manual optimization by automatically searching and measuring performance across many possible configurations.
Solution Approach 2:
The system replaces manual operator intelligence with machine learning algorithms. Instead of relying on human operators to intuitively determine optimal processing orders, the patent substitutes automated reinforcement learning that can systematically evaluate and learn optimal sequences through feedback, overcoming human cognitive limitations in solving complex optimization problems.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
The machine learning device effectively reduces cycle time and optimizes processing conditions, enabling efficient teaching of optimal solutions for laser processing robots, even in multi-robot systems, by learning and adapting processing orders.
Implementation Method 1
a state observation unit that observes, as a state variable, one of a plasma light from a laser processing point of the laser processing robot
Implementation Method 2
a state observation unit that observes, as a state variable, one of a plasma light from a laser processing point of the laser processing robot and a processing sound from the laser processing point of the laser processing robot
Data Source
AI summary
A machine device for learning a processing order of a laser processing robot, includes a state observation unit that observes, as a state variable, one of a plasma light from a laser processing point of the laser processing robot and a processing sound from the laser processing point of the laser processing robot; a determination data obtaining unit that receives, as determination data, a cycle time in which the laser processing robot completes processing; and a learning unit that learns the processing order of the laser processing robot based on an output of the state observation unit and an output of the determination data obtaining unit.