Robotic Object Detection Using Minimum Viable Regions for Unknown Loads

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing robotic systems face challenges in efficiently detecting and handling unknown or unrecognized objects, leading to potential mishandling and increased operational disruptions, especially when objects are heavy or irregularly shaped, as they rely on pre-registered data and may fail to accurately identify physical characteristics.

Innovation Solution

A robotic system equipped with autonomous object detection and registration mechanisms that utilize image comparison and sensor data to identify unrecognized objects by deriving minimum viable regions (MVRs) for gripping and lifting, allowing for autonomous registration and improved object recognition through updated data processing.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If the robotic system relies on pre-registered data for object detection, then the system structure remains simple, but the system fails to accurately identify unrecognized objects leading to mishandling

Engineering Contradiction:
Improveobject recognition accuracyVSAvoiddetection mechanism complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The system performs preliminary actions by capturing images of objects before handling and deriving Minimum Viable Regions (MVRs) in advance. This allows the robotic system to prepare grip locations and lifting points before actual manipulation, improving recognition accuracy for unrecognized objects without requiring complex real-time analysis during handling

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system introduces an intermediary processing layer that compares captured images against a database of registered objects. When matches are found, physical characteristics are retrieved; when no match is found, MVR derivation serves as an intermediary method to still enable safe handling. This intermediary approach bridges the gap between simple pre-registered data and accurate object identification

Inventive Principle:
Principle #24Intermediary (Mediator)

2Reliability

If the robotic system uses manual registration for unrecognized objects, then object handling accuracy improves, but operational time and productivity decrease

Engineering Contradiction:
Improveobject handling reliabilityVSAvoidoperational throughput
Core Design Contradiction:
ReliabilityVSProductivity

Solution Approach 1:

The system enables self-service by automatically deriving MVRs from captured images when objects are unrecognized. Instead of requiring manual registration for every new object type, the system autonomously determines grip locations and lifting points by analyzing image data and identifying geometric features, maintaining reliability while preserving productivity

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The system implements feedback by comparing captured images with the registered objects database and using the results to guide subsequent actions. When objects are successfully matched, the system retrieves accurate physical characteristics; when unmatched, it applies MVR derivation as a fallback feedback mechanism. This feedback loop ensures reliable handling without requiring manual intervention for each object

Inventive Principle:
Principle #23Feedback

3Extent of automation

If the robotic system derives MVRs from captured images, then autonomous handling of unrecognized objects is enabled, but the complexity of image processing increases

Engineering Contradiction:
Improveautonomous object detectionVSAvoidimage processing complexity
Core Design Contradiction:
Extent of automationVSDevice complexity

Solution Approach 1:

The system applies segmentation by dividing the image processing task into distinct stages: first comparing captured images against the registered objects database to identify matches, then separately deriving MVRs from image data only when no match is found. This segmentation of the detection process reduces overall complexity by handling the common case (recognized objects) through simple comparison and reserving complex MVR derivation for the less common case of unrecognized objects

Inventive Principle:
Principle #1Segmentation

Data Source

PatentUS11797926B2Robotic system with automated object detection mechanism and methods of operating the same
Publication Date: 2023.10.24 MUJIN INC
  • US11797926B2 patent drawing
  • US11797926B2 patent drawing
  • US11797926B2 patent drawing

AI summary

A system and method for operating a robotic system to register unrecognized objects is disclosed. The robotic system may use first image data representative of an unrecognized object to derive an initial minimum viable region (MVR). The robotic system may analyze second image data representative of the unrecognized object to detect a condition representative of an accuracy of the initial MVR. The robotic system may register the initial MVR or an adjustment thereof based on the detected condition.