Suction Pad Grasp Planning Using 3D Convolution Kernels
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional robot systems for object handling require re-learning when the number or position of suction pads changes, making it difficult to determine an appropriate handling operation without re-training.
Innovation Solution
A handling device with a manipulator, normal grid generation unit, hand kernel generation unit, and control unit that calculates the ease of grasping using 3D convolution based on spatial data and hand kernels, allowing for adaptive grasping operations without re-learning.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If the number or position of suction pads is changed, then the handling capability is improved, but re-learning is required which reduces productivity
Solution Approach 1:
The patent replaces the conventional machine learning-based approach with a physics-based calculation system. Instead of using neural networks that require re-training, the system uses suction force calculations based on physical principles (suction pad area, pressure, object weight) to determine grasping positions. This substitution eliminates the need for re-learning when suction pad configurations change.
Solution Approach 2:
The system dynamically adjusts calculation parameters based on the actual suction pad configuration. When the number or position of suction pads changes, the system updates the suction force distribution parameters and recalculates grasping positions using the same fundamental algorithm, avoiding the need for re-learning while adapting to new configurations.
2Measurement precision
If conventional machine learning methods are used, then grasping accuracy is achieved, but system complexity increases due to re-learning requirements
Solution Approach 1:
The patent replaces complex machine learning systems with a simpler physics-based calculation system. The grasping determination is achieved through direct calculation of suction forces based on physical parameters, eliminating the need for neural networks, training data, and complex inference processes while maintaining grasping accuracy.
Solution Approach 2:
The system determines grasping positions through self-contained physical calculations rather than external machine learning models. The calculation unit uses suction pad configuration data and object parameters to directly compute optimal grasping positions without requiring external training data or complex inference systems.
Data Source
AI summary
A handling device according to an embodiment includes a manipulator, a normal grid generation unit, a hand kernel generation unit, a calculation unit, and a control unit. The normal grid generation unit converts a depth image into a point cloud, generates spatial data including an object to be grasped that is divided into a plurality of grids from the point cloud, and calculates a normal vector of the point cloud included in the grid using spherical coordinates. The hand kernel generation unit generates a hand kernel of each suction pad. The calculation unit calculates ease of grasping the object to be grasped by a plurality of suction pads based on a 3D convolution calculation using a grid including the spatial data and the hand kernel. The control unit controls a grasping operation of the manipulator based on the ease of grasping the object to be grasped by the plurality of suction pads.