Robot Suction Grasp Control in Tightly Packed Storage
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing model-based approaches for robot systems fail to effectively grasp objects in tightly packed environments due to interference from other objects, as they assume objects are in free space and do not segment objects correctly, leading to ineffective grasp points for suction cup end effectors.
Innovation Solution
A control device and method that calculates grasp points based on the properties of the object's surface, using organized point clouds to determine normals and principal curvatures, allowing the robot system to grasp objects in densely packed conditions by identifying suitable grasp points without relying on traditional model-based approaches.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If model-based approach is used with database storing object models, then object recognition accuracy is improved, but it becomes impractical when there is a large number of object types or wide variations in appearance
Solution Approach 1:
The patent extracts the essential geometric features (normals and curvatures) from the point cloud data directly, eliminating the need for storing comprehensive object models in a database. This allows the system to work with raw sensor data without requiring extensive pre-stored models for each object type.
Solution Approach 2:
The patent changes the approach from storing complete object models to storing and processing only critical geometric parameters (normals and principal curvatures) at each point in the point cloud. This parameter transformation enables the system to handle diverse object types without increasing database complexity.
2Device complexity
If parallel jaw grippers are used for model-free picking, then no database model is needed, but the approach does not segment objects correctly and multiple objects may be detected as a single object
Solution Approach 1:
The patent applies segmentation by dividing the point cloud into distinct object regions based on geometric continuity and curvature characteristics. This allows individual objects to be identified and separated even when tightly packed, enabling accurate grasp point calculation for each object independently.
3Extent of automation
If parallel jaw grippers are used, then model-free operation is achieved, but the grippers do not permit operation in tightly packed environments due to interference with other objects
Solution Approach 1:
The patent calculates grasp points based on local surface properties (normals and curvatures) at specific locations on each object, rather than requiring global gripper positioning. This local approach enables the suction cup end effector to grasp objects in tightly packed environments by focusing on the immediate grasp region without interference from surrounding objects.
4Speed
If grasp points are calculated without considering surface properties, then calculation speed is improved, but the grasp points do not provide effective positions for suction cup end effector
Solution Approach 1:
The patent performs preliminary calculation of surface normals and principal curvatures for all points in the point cloud before selecting grasp points. This pre-computation of geometric properties enables rapid identification of suitable grasp locations without requiring complex iterative optimization, thus maintaining calculation speed while ensuring grasp effectiveness.
Data Source
AI summary
A control device is provided to reliably locate objects and calculate appropriate grasp points for each object to thereby effectively control a robot system. grasp point calculation is based on, at least, properties of a surface of the object to be grasped by the robot system to more effectively calculate a grasp point for the surface of the object. An organised point cloud generator and a robot system having a suction cup end are arranged to generate an organised point cloud of a storage. The robot system can grasp the object from the storage means. Normals of the organised point cloud and principal curvatures of the organised point cloud can be calculated. A grasp point can be calculated for the suction cup end effector of the robot system to grasp an object based on the organised point cloud, the calculated normal, the calculated principal curvatures, and a normal of a lower surface of the storage.