Visual Robot Arm Control Without Manual Camera Calibration

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing methods for visually controlling robot arms require manual calibration, which is time-intensive and labor-intensive, and cannot accurately determine how to move the robot arm to reach a target point without precise initial camera positioning.

Innovation Solution

A method for visually controlling a robot arm that involves placing a camera to capture the target point, determining a vector connecting the reference point to the target, breaking down this vector into standard displacements, and using weighting factors to move the robot arm in a predetermined standard distance in each degree of freedom, allowing for precise control without manual calibration.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Manufacturing precision

If manual calibration is performed to enable precise robot arm control, then manufacturing precision is improved, but loss of time increases due to the time-intensive nature of manual calibration

Engineering Contradiction:
Improverobot arm positioning precisionVSAvoidcalibration time
Core Design Contradiction:
Manufacturing precisionVSLoss of time

Solution Approach 1:

The system performs self-calibration by automatically determining the relationship between camera coordinates and robot arm coordinates through a series of automated measurements and calculations, eliminating the need for manual calibration operations while achieving precise positioning control

Inventive Principle:
Principle #25Self-service

2Manufacturing precision

If manual calibration is performed by highly skilled labor, then manufacturing precision is improved, but quantity of substance increases due to the large amount of skilled labor required

Engineering Contradiction:
Improverobot arm positioning precisionVSAvoidskilled labor quantity
Core Design Contradiction:
Manufacturing precisionVSQuantity of substance

Solution Approach 1:

The calibration process is automated through the controller which performs all calibration calculations and coordinate transformations automatically, completely eliminating the need for highly skilled manual labor while maintaining high positioning precision

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The manual mechanical calibration process is replaced by an automated computational system that uses camera images and mathematical algorithms to determine the coordinate relationship, substituting human skill with automated image processing and calculation

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

3Ease of operation

If the camera position is not precisely defined in advance, then ease of operation is improved, but measurement precision deteriorates because the controller cannot determine how to move the robot arm to reach the target point

Engineering Contradiction:
Improvecamera positioning flexibilityVSAvoidtarget point positioning accuracy
Core Design Contradiction:
Ease of operationVSMeasurement precision

Solution Approach 1:

The system performs preliminary calibration measurements by capturing images at multiple robot arm positions and calculating the coordinate transformation relationship before actual operation begins, enabling flexible camera positioning while ensuring subsequent measurement precision

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system uses feedback from camera images to continuously determine the robot arm's actual position and calculate the required movements to reach target points, maintaining high positioning accuracy regardless of the initial camera position definition

Inventive Principle:
Principle #23Feedback

Data Source

PatentEP3781363B1Method for controlling a robot arm
Publication Date: 2023.06.07 ABB (SCHWEIZ) AG
  • EP3781363B1 patent drawingFigure 1~3
  • EP3781363B1 patent drawingFigure 4
  • EP3781363B1 patent drawing

AI summary

A method for visually controlling a robot arm (1) which is displaceable in a plurality of degrees of freedom, the robot arm (1) carrying at least one displaceable reference point (14), comprises the steps of a) placing at least one camera (2) so that a target point (13) where the reference point (14) is to be placed is contained in an image output by the camera (2); b) displacing the robot arm (1) so that the reference point (14) is within the image; c) determining a vector (A') which, in the image, connects the reference point (14) to the target point (13); d) choosing (S1, S4) one of the degrees of freedom, moving (S2, S5) the robot arm (1) by a predetermined standard distance (ϕaz, ϕbxy) in said degree of freedom and recording (S3, S6) a standard displacement (Daz, Dbxy) of the reference point (14) within the image resulting from the movement of the robot arm (1); e) repeating step d) at least until the vector (A') can be decomposed (S9) into sum of standard displacements (Daz, Dbxy) times associated weighting factors (Caz, Cbxy); f) moving (S9) the robot arm (1) by the weighting factors times the standard distances (Caz*Daz + Cbxy*Dbxy) associated to the standard displacements of the sum.