Robot Navigation Learning From Stuck-Situation Context Images

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing robotic systems face challenges in autonomously navigating and avoiding obstacles, particularly when they become stuck due to external elements like floor materials or structures, leading to inefficient and unsafe movement.

Innovation Solution

An electronic apparatus equipped with sensors and a processor that captures surrounding images, generates context data, and uses machine learning models to recognize and learn from unable-to-move situations, allowing the robot to predict and avoid such situations by adjusting its path.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If the robot uses traditional obstacle avoidance methods, then it can navigate simple environments, but it becomes stuck when encountering complex floor materials or structures

Engineering Contradiction:
Improverobot movement reliabilityVSAvoidadaptability to different floor conditions
Core Design Contradiction:
ReliabilityVSAdaptability or versatility

Solution Approach 1:

The system captures images and generates context data about the environment before the robot actually encounters obstacles. By pre-processing environmental information and identifying potential stuck situations in advance, the robot can plan its path to avoid these situations rather than reacting when stuck

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system uses captured images and context data to continuously monitor the robot's movement status and detect when it is stuck. This feedback loop allows the robot to recognize stuck situations, learn from them through machine learning models, and adjust its navigation strategy accordingly

Inventive Principle:
Principle #23Feedback

2Measurement precision

If the robot captures and processes surrounding images continuously, then it can learn from stuck situations, but the computational complexity and processing time increase

Engineering Contradiction:
Improvesituation recognition accuracyVSAvoidprocessing system complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The system extracts only the necessary context data from captured images - specifically focusing on features relevant to stuck situations - rather than processing entire images. This extraction approach reduces computational complexity while maintaining the ability to accurately recognize and learn from stuck situations

Inventive Principle:
Principle #2Taking out (Extraction)

Data Source

PatentEP3605409B1Electronic device and operation method therefor
Publication Date: 2021.10.20 SAMSUNG ELECTRONICS CO LTD
  • EP3605409B1 patent drawingFigure 1
  • EP3605409B1 patent drawingFigure 2
  • EP3605409B1 patent drawingFigure 3

AI summary

An electronic apparatus and an operating method are provided. The electronic apparatus includes a storage, at least one sensor, and at least one processor configured to execute stored instructions to while the electronic apparatus is moving, capture a surrounding image by using the at least one sensor, when an unable-to-move situation occurs while the electronic apparatus is moving, generate context data including a surrounding image captured within a predetermined time from a time when the unable-to-move situation has occurred, store, in the storage, the generated context data corresponding to the unable-to-move situation having occurred, and learn the stored context data by using one or more data recognition models.