Map Data Correction for Loop Closing in Autonomous Navigation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing SLAM techniques for autonomous moving objects, such as AGVs, struggle with achieving highly accurate position and orientation estimation due to insufficient accuracy in calculating the relative position and orientation of map elements, especially when loop closing detection involves measurement points that are close to each other in real space.
Innovation Solution
An information processing apparatus is designed to acquire three-dimensional map information, detect repeated visits to certain locations, select relevant map elements, and correct the position and orientation information by calculating relative positions and orientations between selected map elements and between the sensor and map elements, thereby improving the accuracy of map data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If loop closing detection is performed only between measurement points that are close to each other in real space, then the processing complexity is reduced, but the accuracy of the calculated relative position and orientation becomes insufficient
Solution Approach 1:
The patent segments the loop closing detection process into two distinct stages: (1) initial detection between close measurement points to identify potential loops, and (2) refined calculation involving multiple map elements from different movement paths to achieve high accuracy. This segmentation allows the system to balance processing efficiency with measurement precision by handling different aspects of the problem at appropriate complexity levels.
Solution Approach 2:
The patent transitions from two-dimensional spatial proximity (close measurement points in real space) to a multi-dimensional approach by incorporating map elements from multiple movement paths and performing calculations across different temporal and spatial dimensions. This enables the system to achieve high accuracy without being constrained by simple spatial proximity.
2Measurement precision
If more map elements are selected for calculation to improve accuracy, then the position and orientation estimation becomes more accurate, but the processing time and computational load increase
Solution Approach 1:
The patent performs preliminary selection of map elements that are likely to contribute to accurate loop closing detection, based on their spatial and temporal relationships with the detected loop. By pre-identifying relevant map elements from multiple movement paths, the system avoids the computational burden of processing all available map elements while still achieving high accuracy through targeted calculations.
Data Source
AI summary
Provided is an information processing method for generating accurate map data by loop closing even when map elements where a loop is detected in a real space are close to each other. The method includes selecting a map elements group including at least one map element on each movement path, the at least one map element including sensor information included in each map element used to detect that the moving object has reached a location near a point a plurality of times, acquiring a relative position and orientation of at least one set of map elements in the map elements group, acquiring a relative position and orientation of the sensor between map elements corresponding to each movement path where the moving object has reached the location near the point the plurality of times, and correcting position and orientation information about the sensor included in each of the map elements.