Vision-Based Robot Tool Attachment for Arbitrary Pose Acquisition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Robots in food-service settings face challenges in manipulating and assembling meals due to unpredictable shape changes of food ingredients, requiring innovative methods to accurately locate and attach tools in unstructured environments.
Innovation Solution
A system that uses image processing and neural networks to determine the pose of tools within the environment, calculating a trajectory for the robot to accurately attach and detach tools without pre-defined positions, employing depth and RGB images, and incorporating tactile feedback for precise mating.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If traditional fixed tool changer interfaces are used, then tool location is predetermined and simple, but robot cannot adapt to tools left in arbitrary positions within the worksite
Solution Approach 1:
The patent replaces the traditional mechanical fixed-position tool changer with a vision-based system using RGB and depth cameras to detect tool poses, substituting mechanical predetermined positioning with optical sensing and computational analysis
Solution Approach 2:
The system enables the robot to autonomously locate and acquire tools by itself without human intervention or predefined positioning, using self-contained vision sensors and onboard computing to determine tool poses and calculate trajectories
2Adaptability or versatility
If vision-based tool detection is implemented, then robot can acquire tools at any pose, but image processing and pose determination complexity increases
Solution Approach 1:
The patent combines RGB imaging and depth sensing into a unified vision system, merging color information with spatial data to achieve accurate 3D pose determination of tools in arbitrary positions and orientations
Solution Approach 2:
The system introduces computer vision algorithms and pose estimation software as intermediaries between the physical tool and the robot's control system, translating visual data into actionable spatial information for trajectory calculation
3Adaptability or versatility
If trajectory calculation is added to enable arbitrary tool attachment, then robot can mate with tools anywhere in workspace, but computational requirements and processing time increase
Solution Approach 1:
The system pre-calculates and stores valid attachment poses and trajectories for common tool types and configurations, allowing rapid retrieval and execution without performing full computational analysis in real-time
Solution Approach 2:
The patent focuses computational resources on determining the specific pose and trajectory for the current tool acquisition task rather than analyzing the entire workspace, optimizing processing by concentrating on local relevant information
Data Source
AI summary
Current technologies allow a robot to acquire a tool only if the tool is in a set known location, such as in a rack. In an embodiment, a method and corresponding system, can determine the previously unknown pose of a tool freely placed in an environment. The method can then calculate a trajectory that allows for a robot to move from its current position to the tool and attach with the tool. In such a way, tools can be located and used by a robot when placed at any location in an environment.