Vision-Guided Robotic Assembly Without Manual Programming
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The deployment of robotic assembly systems for new parts is time-consuming and requires expert technicians, making it financially unviable for small projects due to the complexity of managing configuration data and software settings across multiple devices in automation systems.
Innovation Solution
A vision-based programless assembly system using machine learning and calibrated vision-guided robots to generate 3D models of products, automatically creating assembly recipes and validation protocols, allowing for easy deployment and zero-downtime changeovers with minimal human input, utilizing off-the-shelf cameras and sensors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If expert technicians manually configure robotic assembly systems for new parts, then assembly precision and reliability are improved, but deployment time and cost increase significantly
Solution Approach 1:
The robotic system performs self-configuration by automatically capturing images of new parts, generating 3D models, identifying features, and creating assembly programs without human intervention. The system trains its own vision models and adjusts parameters autonomously, eliminating the need for expert technicians to manually program each new assembly task.
Solution Approach 2:
The system performs preliminary actions by pre-capturing images of new parts before actual assembly begins, generating 3D models and feature databases in advance. This allows the robotic system to be pre-configured with all necessary geometric and spatial information, enabling rapid deployment without time-consuming on-site programming.
2Manufacturing precision
If expert technicians manually program assembly recipes for new products, then assembly precision is improved, but the complexity of managing configuration data increases
Solution Approach 1:
The patent replaces manual mechanical programming with an automated vision-based system. Instead of technicians manually configuring robots, the system uses image capture, 3D modeling, and machine learning to automatically generate assembly programs. This substitution eliminates the complexity of manual configuration data management while maintaining high precision through automated feature recognition and parameter optimization.
Solution Approach 2:
The system creates accurate digital copies (3D models) of physical parts by capturing images and generating detailed geometric representations. These digital twins contain all necessary measurement and positioning information, replacing complex manual configuration data with simplified visual models that the robotic system can directly interpret and execute.
3Productivity
If traditional robotic systems are deployed for small projects, then automation benefits are achieved, but the cost becomes financially unviable due to extensive setup requirements
Solution Approach 1:
The robotic system serves itself by autonomously configuring for new assembly tasks without requiring expert technicians. This self-configuration capability dramatically reduces deployment costs and time, making automated robotic assembly economically viable for small production runs and custom projects where manual programming previously made automation prohibitively expensive.
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
Enables efficient and cost-effective deployment of robotic assembly systems by automating the configuration process, reducing the need for extensive manual setup and expert guidance, and allowing for seamless integration of new products with the same end-of-arm tools and hardware.
Implementation Method 1
The system uses vision in combination with photogrammetry and computer aided design (CAD) data for the product
Implementation Method 2
A vision-based programless assembly system using machine learning and calibrated vision-guided robots to generate 3D models of products
Data Source
AI summary
A method of programless assembly of a device using a robotic cell comprising calibrating a robotic cell for generating a product model, the robotic cell including an end-of-arm camera and utilizing the robotic cell to generate the product model of the device for assembly. The method in one embodiment further comprises validating the product model using a partial assembly method and making the product model available for use by robotic cells.