Controlling method for artificial intelligence moving robot
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing moving robots face challenges in accurately recognizing their position in dynamic environments due to environmental changes such as variations in illumination and object positions, which limits their ability to navigate effectively and efficiently.
Innovation Solution
The method involves generating and using a region-based map by acquiring images through an image acquisition unit, extracting region feature information, and determining the current position based on SLAM-based node information and extracted feature information, allowing for more accurate position estimation and recognition.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If feature points are extracted from images for position recognition, then the robot can recognize position based on visual information, but the accuracy is affected by environmental changes such as illumination variations and object position changes
Solution Approach 1:
The patent segments the image into multiple regions and extracts feature information from each region separately. By dividing the image into regions and processing them independently, the system can identify stable features that remain consistent across different environmental conditions, thereby improving position recognition reliability while maintaining accuracy.
Solution Approach 2:
The patent changes the parameters used for feature extraction by considering multiple region features rather than relying on a single feature point. By varying the feature extraction approach across different regions and selecting features that show minimal change under environmental variations, the system achieves more reliable and accurate position recognition.
2Measurement precision
If sensors such as laser sensors and ultrasonic sensors are used for position recognition, then the robot can recognize position in kidnapping situations, but the cost is greatly increased
Solution Approach 1:
The patent creates a virtual map copy of the environment using visual information from cameras and feature extraction. This virtual representation allows the robot to recognize its position without requiring additional physical sensors like laser or ultrasonic sensors, thereby achieving kidnapping situation recognition while avoiding increased hardware cost and complexity.
Solution Approach 2:
The patent replaces mechanical sensor-based position recognition systems with a vision-based computational approach. By substituting physical sensors (laser, ultrasonic) with image processing and feature matching algorithms, the system achieves the same functional capability of position recognition in kidnapping situations without the associated cost and complexity increases.
3Measurement precision
If the robot uses traveling information for continuous position estimation, then it can maintain position awareness during movement, but it cannot recognize unknown current position when position is forcibly changed by external factors
Solution Approach 1:
The patent implements a feedback mechanism where the robot continuously compares its expected position (based on traveling information) with the actual position determined through feature matching against the virtual map. This feedback loop allows the system to detect position discrepancies caused by kidnapping situations and correct them by re-aligning with the map features, thereby maintaining both continuous position awareness and adaptability to external position changes.
Solution Approach 2:
The patent performs preliminary actions by pre-building a virtual map of the environment and pre-identifying feature points before the robot encounters kidnapping situations. This preparation enables the robot to quickly recognize its actual position after being moved, as the feature matching process can immediately compare current visual information against the pre-established virtual map, combining continuous estimation with kidnapping recovery capability.
Data Source
AI summary
A controlling method for an artificial intelligence moving robot according to an aspect of the present disclosure includes: moving based on a map including a plurality of regions; acquiring images from the plurality of regions through an image acquisition unit during the moving; extracting region feature information based on the acquired image; and storing the extracted region feature information in connection with position information when an image is acquired.