Robot Control Method for Accurate Assembly via Visual Feedback

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing robot control methods often fail to accurately assemble components due to unintended deviations in the relative position between the robot hand and the target object, leading to incorrect assembly even when the robot follows a taught trajectory.

Innovation Solution

A computer-implemented method that acquires a current image of the positional relationship between objects, determines if they are in a reference positional relationship, and controls the robot's position and posture to achieve a target positional relationship by comparing the current image with a reference image, ensuring accurate assembly even with gripping displacement.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If robot control is performed based on taught trajectory alone, then the robot can follow a predetermined path, but unintended deviation occurs in the relative position between the robot hand and target object leading to incorrect assembly

Engineering Contradiction:
Improveassembly accuracyVSAvoidpositional relationship measurement
Core Design Contradiction:
ReliabilityVSMeasurement precision

Solution Approach 1:

The patent implements visual feedback by capturing images with a camera to detect the actual positional relationship between the robot hand and target object during assembly. This feedback information is used to correct deviations from the taught trajectory, ensuring accurate assembly even when gripping displacement or positioning errors occur.

Inventive Principle:
Principle #23Feedback

Solution Approach 2:

The patent performs preliminary teaching of the assembly trajectory and stores reference images showing the correct positional relationship before actual assembly operations. These pre-established references are then used during execution to compare against real-time visual feedback and maintain assembly accuracy.

Inventive Principle:
Principle #10Preliminary action

2Reliability

If visual servoing is used to measure position change, then feedback information is obtained for control, but the system complexity increases when incorporating multiple control systems

Engineering Contradiction:
Improveposition control accuracyVSAvoidcontrol system structure
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent merges visual servoing control with force control into a unified control system. By integrating these control methods and managing them through a single control device, the system achieves accurate position control while avoiding the complexity of multiple separate control systems.

Inventive Principle:
Principle #5Merging (Combining)

3Reliability

If force control is combined with visual servoing, then both load information and positional relationship are utilized, but the control system becomes more complex

Engineering Contradiction:
Improveassembly control accuracyVSAvoidcontrol system integration
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent combines force control and visual servoing control outputs through a unified control algorithm that processes both load information and positional relationship data. This integration is managed within a single control device, achieving improved assembly control accuracy while maintaining manageable system complexity.

Inventive Principle:
Principle #5Merging (Combining)

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

This method ensures accurate assembly by confirming the positional relationship through visual feedback and adjusting the robot's position and posture, preventing incorrect assembly and potential damage from collisions.

Implementation Method 1

acquiring a current image showing a positional relationship between a first object held by the robot and a second object, determining whether the first object and the second object are in a reference positional relationship based on the current image

Methodology Applied
Scientific EffectVisual feedback: Photography

Data Source

PatentUS20250100154A1Computer-implemented method for controlling a robot, robot control method, system, article manufacturing method, and recording medium
Publication Date: 2025.03.27 CANON KK
  • US20250100154A1 patent drawing
  • US20250100154A1 patent drawing
  • US20250100154A1 patent drawing

AI summary

A computer-implemented method for controlling a robot includes acquiring a current image showing a positional relationship between a first object held by the robot and a second object, determining whether the first object and the second object are in a reference positional relationship based on the current image, and controlling, in a case where a computer determines that the first object and the second object are in a reference positional relationship based on the current image, a position and a posture of the robot according to a trajectory of the robot set before the determining such that the first object and the second object have a target positional relationship.