AI Welding Object Detection for Adaptive Robotic Path Planning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In heavy industries, robotic welding processes face challenges in determining the optimal welding path due to complex and varying tasks, where specific welding objects and their poses are not known, requiring an efficient method for online sensing and detection.
Innovation Solution
A method utilizing artificial intelligence, such as deep learning and neural networks, to automatically detect and identify welding objects in a welding environment by obtaining scanning data, determining their pose, and generating a welding path, leveraging a combination of real and simulated 3D data for training the AI algorithm.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If traditional pre-programmed welding paths are used, then welding precision is maintained, but adaptability to complex and varying welding tasks deteriorates
Solution Approach 1:
The system performs preliminary scanning and detection of the welding environment to create a digital model before welding begins. This preliminary action enables the robot to adapt to complex and varying welding tasks while maintaining precision through AI-based path planning rather than relying on pre-programmed paths
Solution Approach 2:
The welding system transitions from static pre-programmed paths to dynamic adaptive path planning. The robot uses real-time scanning data and AI algorithms to generate and adjust welding paths dynamically based on the actual welding environment and detected objects
2Adaptability or versatility
If AI-based object detection is implemented, then adaptability to unknown welding objects improves, but system complexity increases
Solution Approach 1:
The system employs a multi-functional integrated approach where a single scanning system performs multiple functions: capturing geometric data, identifying welding objects, determining poses, and generating welding paths. This universal system handles various welding objects and tasks without requiring separate specialized systems for each function
Solution Approach 2:
The patent introduces AI algorithms and software intermediaries that bridge the gap between raw scanning data and welding execution. These intermediaries process and interpret sensor data, enabling the robot to understand and adapt to unknown welding objects without increasing hardware complexity
3Adaptability or versatility
If online sensing and detection are performed, then adaptability to varying tasks improves, but measurement and detection difficulty increases
Solution Approach 1:
The system implements online feedback through continuous scanning and detection during the welding process. The AI algorithms analyze detected objects and their poses in real-time, providing feedback that enables dynamic adjustment of welding paths and parameters to adapt to varying tasks
Solution Approach 2:
The patent creates digital copies or models of the welding environment and objects through scanning. These digital representations simplify the detection and measurement process by allowing AI algorithms to work with processed model data rather than raw sensor data, reducing detection difficulty
Data Source
AI summary
A system and a method is for automating welding processes, in particular welding processes in the heavy industries. One embodiment regards a computer implemented method for automatic detection and/or planning of a welding task in a welding environment, the method including the steps of: obtaining scanning data from a scan of the welding environment, detecting welding object(s) in the scanning data by means of artificial intelligence employing a machine learning algorithm, wherein the machine learning algorithm has been trained on real and simulated 3D data of known welding objects, determining the pose of each detected welding object, and optionally generating a welding path for each detected welding object.