Robot Gripper Control for Irregular Soft Object Handling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Robot systems struggle to effectively grip objects with low reaction forces and irregular shapes, such as cream puffs and daifuku, due to variations in shape and timing differences between movable claws, leading to potential damage and reduced production efficiency.
Innovation Solution
A robot system equipped with a machine learning device that utilizes stop reference data, distance data, and comparison data to construct a model for optimizing the operation mode, including positioning and gripping operations, by adjusting the positions of movable claws based on machine learning algorithms.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a pair of movable claws pinches an object to grip the object, then the object can be held securely, but one of the movable claws may come into contact with the object before the other, causing local high pressure or sliding
Solution Approach 1:
The robot system performs a positioning operation before the gripping operation to adjust the grip unit to a predetermined position where the object is located between the pair of movable claws. This preliminary positioning ensures that both claws are correctly aligned with the object before pinching begins, preventing one claw from contacting the object before the other and avoiding local high pressure or sliding damage.
2Device complexity
If the robot system uses a fixed reference value for stopping the gripping operation, then the control is simple, but it cannot adapt to objects with varying shapes and positions
Solution Approach 1:
The reference value for stopping the gripping operation is made variable rather than fixed. The controller adjusts the reference value dynamically based on the detected position and shape characteristics of the object. This allows the system to adapt to objects with varying shapes and positions while maintaining effective grip control, resolving the contradiction between control simplicity and adaptability.
3Manufacturing precision
If the robot system performs positioning operation to adjust claw positions, then gripping precision is improved, but the operation time increases
Solution Approach 1:
The robot system uses feedback from object detection (image processing or other sensing) to determine the optimal grip position and adjust the positioning operation accordingly. The feedback mechanism allows the system to quickly identify the correct positioning parameters based on object characteristics, minimizing the time required for positioning while maintaining high gripping precision for objects with irregular shapes.
Data Source
AI summary
In a robot (industrial robot) system, a robot holds a workpiece by pinching the workpiece between movable claws. A controller, which controls the robot, includes a host controller that controls the robot to perform a positioning operation for positioning the hand to a grip position and a gripping operation for displacing each of the movable claws toward each other at the grip position. In the controller, a machine learning device acquires stop reference data set for gripping of the workpiece, distance data indicating a distance between each of the movable claws of the hand positioned at the grip position and the workpiece, and comparison data indicating a deformation amount of the workpiece before and after the gripping operation. The machine learning device performs machine learning using such acquired data, resulting in constructing a model used for setting an operation mode of the gripping operation.