Robotic Object Handling Using Pickable Region Surface Maps
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Robots lack the sophistication to duplicate human interactions required for executing complex tasks, particularly in identifying and handling objects with irregular arrangements, such as boxes and pouches, which are challenging due to the difficulty in identifying suitable pickable regions.
Innovation Solution
A computing system that communicates with a robot arm and a camera to generate a surface cost map, segment pickable regions, and create a motion plan for transferring objects based on these regions, using image information to improve object handling precision and accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If robots are used to execute tasks in manufacturing and packaging, then productivity is improved, but the ability to handle objects with irregular arrangements deteriorates
Solution Approach 1:
The system segments the object handling task into multiple stages: image acquisition, point cloud generation, pickable region identification, and motion planning. This segmentation allows each stage to be optimized independently, enabling the robot to handle irregularly arranged objects effectively while maintaining high productivity
Solution Approach 2:
The system introduces an intermediary computational layer between the robot and the objects, using cameras to capture images and generate point clouds, and using algorithms to identify pickable regions. This intermediary processing enables the robot to adapt to irregular object arrangements without sacrificing speed
2Adaptability or versatility
If sophisticated object detection and handling techniques are implemented, then the ability to handle irregular objects is improved, but device complexity increases
Solution Approach 1:
The system uses a multi-functional approach where a single integrated pipeline handles multiple tasks: image capture, point cloud generation, pickable region identification, and motion planning. This universal system reduces overall device complexity compared to having separate specialized systems for each function
Solution Approach 2:
The system employs self-service mechanisms where the robot autonomously identifies pickable regions and generates motion plans without human intervention. The automated identification of suitable grasping points on irregular objects eliminates the need for complex manual programming or external assistance
3Measurement precision
If pickable regions are identified for objects with irregular arrangements, then object handling precision is improved, but measurement and detection difficulty increases
Solution Approach 1:
The system replaces mechanical measurement methods with optical sensing using cameras. Images are captured and converted into point clouds, which are then processed to identify pickable regions. This substitution of optical fields for mechanical measurement enables high precision detection of irregular objects while reducing the difficulty of identification
Solution Approach 2:
The system transitions from 2D images to 3D point clouds, adding a dimensional aspect to object detection. This dimensional transformation enables more accurate identification of pickable regions on irregularly shaped objects by providing depth information and spatial context
Data Source
AI summary
A computing system configured for object transfer is provided. The computing system includes at least one processing circuit configured to identify pickable regions of objects according to image information of the objects. Pickable regions may be determined according to a surface cost map indicating smoothness of regions of the image information, determined according to height differences and normal differences. Identification of pickable regions may be used to in a motion planning operation to transfer the objects.