Multi-Suction Robot Grasp Planning Without Re-Learning

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Conventional robot systems for object handling, such as picking automation, require re-learning when the number or position of suction pads changes, making it difficult to determine appropriate handling operations without re-learning teacher data.

Innovation Solution

A handling device with a controller that includes a processing unit, planning unit, and control unit, utilizing 3D convolution calculations and hand kernels to determine the ease of grasping and optimal posture for multiple suction pads, allowing for adaptive handling operations without re-learning.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If the number or position of suction pads is changed, then the handling device can adapt to different object sizes and shapes, but re-learning using teacher data is required which reduces productivity

Engineering Contradiction:

Improveadaptability to different object sizes and shapesVSAvoidproductivity reduced by re-learning requirement

Core Design Contradiction:

Adaptability or versatilityVSProductivity

Solution Approach 1:

The handling device performs self-adjustment by automatically calculating modified teacher data based on the actual suction pad configuration. The system uses the originally learned teacher data and computationally derives new training data that reflects the current suction pad arrangement, eliminating the need for external re-learning interventions and enabling continuous adaptive operation.

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The system pre-calculates and stores modified teacher data for various suction pad configurations before actual handling operations are needed. By preparing the adjusted training data in advance based on predicted or potential suction pad arrangements, the system ensures immediate adaptability without interruption to productivity when configuration changes occur.

Inventive Principle:

Principle #10Preliminary action

2Measurement precision

If conventional robot systems use re-learning with teacher data, then accurate grasping can be achieved, but the process becomes time-consuming and inefficient when configuration changes occur

Engineering Contradiction:

Improvegrasping accuracyVSAvoidtime lost to re-learning process

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The system replaces the time-consuming mechanical re-learning process with a computational data transformation approach. Instead of physically re-training the robot system through repeated trials and errors, the system uses algorithmic processing to modify existing teacher data, rapidly generating accurate grasping parameters for new suction pad configurations without time loss.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

3Productivity

If the handling device uses multiple suction pads, then it can handle larger or irregularly shaped objects, but determining appropriate handling operations becomes more complex

Engineering Contradiction:

Improvehandling capability for larger and irregular objectsVSAvoidcomplexity in determining handling operations

Core Design Contradiction:

ProductivityVSDevice complexity

Solution Approach 1:

The system segments the complex handling problem into manageable components by processing each suction pad's contribution independently. The teacher data is divided and adjusted according to individual suction pad positions and characteristics, allowing the system to systematically determine handling operations for multiple pads without being overwhelmed by overall complexity.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The handling determination process focuses on local characteristics of each suction pad rather than treating the entire multi-pad system as a single complex unit. By analyzing and adjusting grasping parameters locally for each suction pad's specific position and function, the system simplifies the overall complexity while maintaining comprehensive handling capability for various object sizes and shapes.

Inventive Principle:

Principle #3Local quality

Data Source

PatentEP3970923A1Handling device and computer program product

Publication Date: 2022.03.23 KK TOSHIBA

Figure 1
Figure 2A~2B
Figure 3

AI summary

A handling device (100) according to an arrangement includes a manipulator (1, 2), a normal grid generation unit (322), a hand kernel generation unit (323), a calculation unit (324), and a control unit (33). The normal grid generation unit (322) 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 (323) generates a hand kernel of each suction pad. The calculation unit (324) calculates ease of grasping the object to be grasped by a plurality of suction pads (6, 6a, 6b) based on a 3D convolution calculation using a grid including the spatial data and the hand kernel. The control unit (33) controls a grasping operation of the manipulator (1, 2) based on the ease of grasping the object to be grasped by the plurality of suction pads (6, 6a, 6b).