Aligning Robot Tool Reference Frame with Camera Measurement System
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional coordinated movement machines, such as industrial robots, face challenges in accurately positioning tools due to the need for extensive calibration and skilled operators, often requiring multiple sensors and external measurements, which is time-consuming and labor-intensive.
Innovation Solution
A method that aligns a coordinated movement machine tool reference frame with a measurement system reference frame by applying linear and rotational movements, measuring camera movements, and determining angular and linear offsets to transform the reference frames without prior calibration, allowing the camera and robot tool to be aligned using visual guidance and recognition software.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If many sensors are mounted in different locations around the robot working area, then the robot tool positioning accuracy is improved, but the device complexity and setup time increase
Solution Approach 1:
The patent extracts the measurement function from multiple distributed sensors and consolidates it into a single camera-based measurement system. Instead of mounting many sensors around the robot workspace, the invention uses one camera to capture images of a learned object, from which all necessary positional and orientational information is derived through image processing and geometric calculations.
Solution Approach 2:
The single camera system performs multiple functions: it measures the position of the robot tool, determines the orientation of the tool, and establishes the transformation between the robot reference frame and the measurement system reference frame. This multi-functional approach replaces the need for multiple specialized sensors.
2Measurement precision
If external measurements are taken using a two-dimensional checkerboard, then the robot tool alignment is improved, but the calibration time and operator skill requirements increase
Solution Approach 1:
The system performs preliminary learning of a reference object in the camera's field of view before robot calibration. This learned object serves as a permanent reference that eliminates the need for repeated checkerboard measurements during each calibration process. The geometric features of the learned object are stored and used for subsequent alignment operations.
Solution Approach 2:
Instead of using a physical checkerboard that requires physical placement and measurement, the system creates a digital copy of a learned object's geometric features from camera images. This virtual reference model can be repeatedly used without physical re-measurement, significantly reducing calibration time.
3Measurement precision
If a highly skilled robot operator is used for calibration, then the robot positioning accuracy is improved, but the ease of operation decreases
Solution Approach 1:
The system performs self-calibration by automatically comparing the robot's reported tool position with the position derived from camera images of the learned object. The transformation between reference frames is calculated automatically through geometric relationships, eliminating the need for manual intervention and expert judgment in the calibration process.
Solution Approach 2:
The system uses camera measurements as feedback to verify and correct the robot's tool position and orientation. By continuously comparing the robot's claimed position with the visually measured position, the system can detect and correct positioning errors without requiring operator expertise.
Data Source
Figure 1
Figure 2
AI summary
A method for aligning a machine tool reference frame with a camera reference frame includes applying at least two linear machine movements and at least two rotational movements to a machine arm. Measurements are taken during these movements. Transforming the camera reference frame to the machine reference frame is based on an angular offset and x, y and z offsets between a machine tool origin and a camera origin based on the measurements taken during the machine movements.