Remote Laser Welding Scanner Compensation for Robot Bending

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current remote laser welding robot systems face challenges in accurately controlling the robot and laser scanner in close cooperation, leading to mechanical bending and shifts in the laser beam irradiation position, making it difficult to weld at the desired location.

Innovation Solution

A remote laser welding robot system with a separate robot controller and laser irradiation controller, where the robot controller provides feedback on drive information to the laser irradiation controller, which uses a learned model from supervised learning to compensate for actual position and posture data, ensuring precise laser beam alignment.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Device complexity

If the robot controller and scanner controller operate independently, then the system structure is simple and easy to control, but the coordination between robot and laser scanner deteriorates, causing delays in laser scanner operation

Engineering Contradiction:
Improvecontroller structureVSAvoidcoordination between robot and laser scanner
Core Design Contradiction:
Device complexityVSReliability

Solution Approach 1:

The patent implements a feedback mechanism where the robot controller transmits drive information (command speed, command position, command posture) to the scanner controller in real-time. The scanner controller uses this feedback to dynamically adjust laser scanner operation, ensuring coordinated movement between robot and scanner while maintaining independent controller structures.

Inventive Principle:
Principle #23Feedback

Solution Approach 2:

The patent introduces an intermediary communication interface between the robot controller and scanner controller. This intermediary transmits drive information and enables coordination without merging the controller structures, allowing independent controllers to work together seamlessly through standardized information exchange protocols.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Device complexity

If command position/posture information is used to control the robot, then the control system is simple, but mechanical bending occurs that is not detected, causing shifts in laser beam irradiation position

Engineering Contradiction:
Improvecontrol systemVSAvoidlaser beam irradiation position
Core Design Contradiction:
Device complexityVSManufacturing precision

Solution Approach 1:

The patent replaces direct mechanical position sensing with a learned model-based prediction system. Instead of using complex mechanical sensors to detect actual position and posture, the system uses a learned model that processes drive information to calculate predicted actual position and posture, compensating for mechanical bending effects through computational modeling rather than mechanical measurement.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The patent creates a virtual copy of the physical system through the learned model. The learned model replicates the relationship between drive information and actual position/posture, allowing the system to predict and compensate for deviations without physically measuring them. This virtual copying enables precision correction through software rather than hardware modifications.

Inventive Principle:
Principle #26Copying

Data Source

PatentUS11400547B2Laser machine
Publication Date: 2022.08.02 FANUC LTD
  • US11400547B2 patent drawing
  • US11400547B2 patent drawing
  • US11400547B2 patent drawing

AI summary

A laser machine includes a scanner configured to irradiate a workpiece with a laser beam, a robot configured to move the scanner, a robot controller configured to control the robot, and a scanner controller configured to control the scanner so as to control an irradiation position of the laser beam. The scanner controller includes a learned model obtained through supervised learning based on training data including, as input data, drive information relating to the robot at times when the scanner is moved in advance in a plurality of directions and speeds, and as correct data, actual position data and actual posture data of the attached scanner at the times. The actual position data and the actual posture data are calculated on the basis of the drive information relating to the robot in the learned model, and a robot movement consideration/calculation unit compensates the irradiation position of the laser beam.