Controller Optimizes Robot Bonding via Machine Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The determination of optimal operating parameters for bonding a radiation plate to an electronic component by a robot is challenging due to variations in heat conductor viscosity, coating amount, temperature, humidity, substrate strength, and electric component type, leading to increased tact time and potential damage, and requires significant operator effort through trial and error.
Innovation Solution
A controller with a machine learning device that collects data on film thickness and tact time by randomly changing operating parameters and performs machine learning to derive optimal parameters for bonding, including a state observation section, determination data acquisition section, and learning section that associates heat conductor state data with operating parameters, allowing for automated determination of appropriate film thickness and reduced tact time.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If the radiation plate is set at a position close to the electric component to shorten tact time, then the bonding speed increases, but the heat conductor may scatter or the electric component may break due to impact
Solution Approach 1:
The robot performs preliminary actions by first moving to a position above the electric component and then gradually approaching the bonding position. This staged approach allows the system to maintain high productivity while preventing impact damage through controlled, progressive movement rather than direct high-speed collision.
Solution Approach 2:
The system dynamically adjusts the robot's movement speed and positioning based on real-time conditions. The robot can vary its approach speed and positioning accuracy depending on the specific bonding situation, allowing optimization of both bonding speed and quality for each individual case rather than using fixed parameters.
2Reliability
If the pressing pressure is set high to ensure proper bonding, then the bonding quality improves, but the electric component or substrate may break
Solution Approach 1:
The system changes pressing pressure parameters dynamically based on the specific electric component being bonded. By adjusting pressure, bonding time, and temperature parameters according to component type and heat conductor state, the system achieves proper bonding quality without exceeding the strength limits of delicate electric components or substrates.
Solution Approach 2:
The robot uses feedback from sensors and machine learning to monitor the bonding process in real-time. When proper bonding is detected or when resistance increases indicating potential damage risk, the system automatically adjusts or terminates the pressing action, ensuring bonding quality while preventing component breakage.
3Productivity
If the pressing time is shortened to increase productivity, then the bonding speed increases, but appropriate film thickness may not be maintained
Solution Approach 1:
The system changes multiple parameters simultaneously including pressing time, pressure, and temperature based on the heat conductor's viscosity and state. This coordinated parameter adjustment allows the system to maintain appropriate film thickness even with reduced pressing time, thereby increasing productivity without sacrificing bonding quality.
Solution Approach 2:
The robot performs preliminary positioning and preparation actions before the actual bonding process. By pre-positioning the radiation plate accurately and preparing the bonding environment in advance, the system can use shorter pressing times while still achieving proper film thickness and bonding quality.
4Productivity
If the robot operation is optimized by teaching to shorten tact time, then the bonding speed increases, but the determination of operating parameters requires enormous operator effort through trial and error
Solution Approach 1:
The robot performs self-learning and self-optimization through machine learning algorithms. Instead of requiring operators to manually teach and adjust parameters through trial and error, the system automatically learns optimal operating parameters from observed bonding outcomes and environmental conditions, thereby increasing productivity while eliminating enormous operator effort.
Solution Approach 2:
The system uses feedback from bonding results and environmental sensors to automatically adjust operating parameters. The machine learning model continuously improves by learning from past bonding experiences and real-time conditions, enabling the robot to optimize its own operation for high productivity without requiring extensive manual programming or trial-and-error adjustment by operators.
Data Source
AI summary
A controller includes a machine learning device that learns an operating parameter for an operation of bonding a radiation plate by a robot. The machine learning device observes operating parameter data and heat conductor state data as state variables that express a current state of an environment. In addition, the machine learning device acquires determination data indicating a propriety determination result of the operation of bonding the radiation plate, and learns the operating parameter in association with the heat conductor state data, using the state variables and the determination data.