Autonomous Device Rotational Imaging for SLAM Loop Closing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous movement devices face challenges in accurately recognizing their return to previous positions due to error accumulation in Simultaneous Localization And Mapping (SLAM) schemes, especially during looped movements, which can lead to incorrect positioning and navigation issues.
Innovation Solution
The autonomous movement device includes an image picker, driver, controller, and memory, where the controller causes the image picker to rotate the device and search for images with a similarity level greater than a predetermined threshold, correcting the map based on found images to improve positioning accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If SLAM scheme is used for map creation and position estimation, then autonomous movement capability is achieved, but error accumulation occurs leading to position recognition failure
Solution Approach 1:
The patent implements a feedback mechanism by capturing images at regular intervals during movement, comparing current images with historical images, and using the comparison results to correct accumulated position estimation errors. The image recognition system provides feedback on whether the device has returned to a previously visited location, allowing the system to adjust and correct its position estimates based on visual evidence rather than relying solely on incremental odometry data that accumulates errors.
2Measurement precision
If loop closing process is implemented to correct position errors, then position accuracy improves, but device complexity increases
Solution Approach 1:
The patent employs periodic action by capturing images at regular time intervals during autonomous movement rather than continuously. This periodic image capture reduces the computational burden and system complexity compared to continuous monitoring, while still enabling the loop closing process to detect when the device returns to previously visited locations and correct position errors accordingly.
3Measurement precision
If continuous image capture is performed for position correction, then positioning accuracy improves, but energy consumption increases
Solution Approach 1:
The system captures images periodically at predetermined time intervals during autonomous movement rather than continuously. This periodic sampling approach reduces energy consumption significantly compared to continuous image capture, while still providing sufficient visual data at key moments to enable loop closing detection and position correction when the device returns to previously visited locations.
Solution Approach 2:
The system performs preliminary action by capturing and storing images at regular intervals during the exploration phase, creating a visual history database before the device needs to return or correct its position. This preliminary image capture allows the device to later compare current visual data with historical data for loop closing detection, reducing the need for intensive real-time processing and energy consumption during critical positioning moments.
Data Source
AI summary
To increase the frequency of executing a loop closing process, and to reduce an accumulated error in a local device position, and the like. A rotational image picker of an autonomous movement device picks up an image while performing a rotational action. An image memory stores information on the picked-up image. A map memory stores a created map. A position estimator estimates the local device position. A similar image searcher searches, from the image memory, the image that has a similarity level of equal to or greater than a predetermined similarity level. A map corrector corrects the map stored in the map memory when the similar image searcher founds the image that has the similarity level of equal to or greater than the predetermined similarity level.