3D Range-Image Robot Picking for Irregular Workpieces
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional robot systems face challenges in selecting a picking position for workpieces, particularly for irregular shapes and workpieces in process, as they require time-consuming CAD model generation and rely on experiential search algorithms, making the process inefficient and prone to errors.
Innovation Solution
A robot system that utilizes a three-dimensional measuring device to generate range images, a display for teaching the picking position, and a machine learning component to select new picking positions based on taught positions and their peripheries, enabling the robot to pick up workpieces more accurately and efficiently without relying on CAD data or extensive user experience.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If CAD data matching method is used to teach picking positions, then the robot can accurately pick up workpieces, but it takes time to teach multiple picking positions and requires CAD model generation
Solution Approach 1:
The patent uses range image data as a copy of the workpiece geometry instead of requiring original CAD models. The system captures actual workpiece geometry through 3D scanning and uses this captured data for picking position determination, eliminating the need for time-consuming CAD model generation and matching processes.
Solution Approach 2:
The patent replaces the mechanical/CAD-based matching system with an image processing-based system. Instead of using CAD models and geometric matching algorithms, the system uses range images and machine learning to directly determine picking positions, substituting complex mechanical modeling with simpler visual recognition.
2Ease of manufacture
If search algorithm based on parameters is used without CAD data, then no CAD model is required, but it requires selecting algorithms based on experience and is not easily adaptable
Solution Approach 1:
The patent implements a machine learning system that automatically learns and adapts to different workpiece types without requiring manual algorithm selection. The system trains on range image data and autonomously determines appropriate picking strategies, eliminating the need for operators to have experience-based knowledge for algorithm selection.
Solution Approach 2:
The patent changes the approach from fixed parameter-based search algorithms to adaptive machine learning models. The system learns optimal parameters and strategies from training data, allowing it to adapt to different workpiece geometries and picking requirements without manual reconfiguration or algorithm selection.
3Reliability
If CAD model is required for picking position selection, then regular workpieces can be handled, but irregular workpieces and workpieces in process cannot be processed
Solution Approach 1:
The patent segments the workpiece handling process into two independent components: workpiece acquisition through 3D scanning and picking position determination through machine learning. This segmentation allows the system to handle any workpiece type that can be scanned, including irregular shapes and workpieces in process, without requiring pre-existing CAD models for each type.
Solution Approach 2:
The patent creates a universal picking system that can handle multiple workpiece types including regular workpieces, irregular workpieces, and workpieces in process. The machine learning model trained on range image data provides a multi-functional solution that replaces multiple specialized CAD-based approaches, enabling the system to adapt to various workpiece geometries without requiring separate CAD models for each type.
Data Source
AI summary
To select a picking position of a workpiece in a simpler method. A robot system includes a three-dimensional measuring device for generating a range image of a plurality of workpieces, a robot having a hand for picking up at least one of the plurality of workpieces, a display part for displaying the range image generated by the three-dimensional measuring device, and a reception part for receiving a teaching of a picking position for picking-up by the hand on the displayed range image. The robot picks up at least one of the plurality of workpieces by the hand on the basis of the taught picking position.