Mistakenly ingested object identifying robot cleaner and controlling method thereof
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Robot cleaners face limitations in real-time object identification accuracy, leading to potential ingestion of objects like jewelry or hazardous items due to restricted calculation resources, resulting in incorrect identification and subsequent ingestion.
Innovation Solution
A robot cleaner system equipped with a shock detection sensor, camera, and processor that captures images before impact, uses multi-scale patch analysis and artificial intelligence models to identify ingested objects, allowing for post-operation analysis in a rest mode to determine the ingested object's identity and prevent further ingestion.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If real-time image analysis is performed during operation, then the robot cleaner can identify objects while moving, but the calculation resources are restricted leading to low identification accuracy
Solution Approach 1:
The robot cleaner captures images in advance during its normal operation before ingestion occurs. These pre-captured images are stored and later analyzed after the robot returns to the base station, allowing complex AI processing to be performed without real-time computational constraints.
Solution Approach 2:
The robot cleaner autonomously performs post-event object identification by itself using its own captured images and onboard processing capabilities, without requiring external assistance during the critical identification phase.
2Measurement precision
If the robot cleaner uses simple real-time detection, then it can operate smoothly, but it fails to identify small objects like jewelry leading to ingestion
Solution Approach 1:
Images are captured continuously before ingestion events occur, creating a historical record that can be analyzed in detail later. This preliminary capture allows sophisticated analysis of small objects without interfering with real-time cleaning operations.
Solution Approach 2:
The system changes the analysis parameters dynamically - using simple real-time detection during cleaning operations for efficiency, and switching to detailed post-event AI analysis when ingestion is detected, thereby optimizing both efficiency and accuracy for different operational phases.
3Reliability
If the robot cleaner ingests objects without verification, then it maintains continuous operation, but valuable or hazardous objects may be mistakenly picked up
Solution Approach 1:
The shock sensor, which detects the harmful event of object ingestion, is converted into a beneficial trigger that initiates the object identification process. The ingestion event itself becomes the signal to retrieve and analyze the captured images, turning a failure mode into a diagnostic opportunity.
Solution Approach 2:
The shock detection sensor acts as an intermediary between the physical ingestion event and the digital image analysis process. It translates the physical impact into an electronic signal that triggers the retrieval and analysis of pre-captured images, bridging the physical and digital domains.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Enhances object identification accuracy by leveraging AI models for post-event analysis, reducing the risk of ingesting valuable or hazardous items and enabling users to retrieve mistakenly ingested objects, thus improving the robot cleaner's operational safety and reliability.
Implementation Method 1
a shock detection sensor configured to detect impact of the object on the robot cleaner
Data Source
AI summary
A robot cleaner includes an intake port, a shock detection sensor, a camera, a memory, and a processor. The memory may include an artificial intelligence model trained to identify an object, and the processor may, based on the object being ingested by the intake port, identify an image obtained within a preset time before the object is ingested, among a plurality of images obtained through the camera and identify the object according to the artificial intelligence model. Thereby, a user may be informed that the robot cleaner has ingested the object.