Local Correction for SLAM Loop Closing in AGV Navigation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing Simultaneous Localization and Mapping (SLAM) techniques, such as ORB-SLAM2, face challenges in accurately estimating the position and orientation of autonomous moving bodies like AGVs due to long processing times for loop closing, leading to errors and deterioration in position/orientation estimation accuracy.
Innovation Solution
Implementing a local correction processing mechanism that quickly corrects the position and orientation of measurement points when a loop is detected, preventing the generation of redundant measurement points and reducing the accumulation of errors, thereby maintaining accurate position/orientation estimation even during prolonged pose graph optimization.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If loop closing correction processing is performed using existing SLAM techniques, then position and orientation estimation accuracy is improved, but processing time increases significantly causing deterioration in real-time performance
Solution Approach 1:
The patent divides the correction processing into two distinct segments: (1) local correction processing that quickly corrects positions and orientations of measurement points in the vicinity of the sensor, and (2) pose graph optimization that performs comprehensive correction of all measurement points. By segmenting the processing, the system can provide immediate local corrections without waiting for the time-consuming global optimization, thus resolving the contradiction between accuracy and real-time performance.
Solution Approach 2:
The patent performs preliminary local correction processing immediately when a loop is detected, before the sensor moves to new positions and before the time-consuming pose graph optimization is completed. This preliminary action ensures that the most critical measurement points are corrected in real-time, preventing error accumulation while the comprehensive optimization runs in the background.
2Productivity
If the sensor continues moving after correction processing starts, then new measurement points are generated, but errors accumulated before correction completion cause position and orientation estimation to vary
Solution Approach 1:
The system performs preliminary local correction processing immediately when a loop is detected, before the sensor continues moving and generates new measurement points. This preliminary correction establishes a stable reference frame that prevents error accumulation, ensuring that subsequent position and orientation estimations remain consistent even as the sensor continues to operate and generate new data.
Solution Approach 2:
The local correction processing acts as a preliminary countermeasure against the error accumulation that would otherwise occur during continuous operation. By correcting the most critical measurement points before the sensor moves on, the system preemptively neutralizes the source of estimation variability, maintaining reliability throughout continuous operation.
3Measurement precision
If comprehensive correction of all measurement points is performed, then overall map accuracy is improved, but processing time increases and redundant measurement points are generated
Solution Approach 1:
The patent segments the correction processing into two phases with different scopes: local correction that quickly processes only the most critical measurement points near the sensor, and pose graph optimization that comprehensively corrects all measurement points. This segmentation allows the system to maintain high processing efficiency through immediate local corrections while still achieving overall map accuracy through the subsequent comprehensive optimization phase.
Solution Approach 2:
The patent implements a two-stage approach where the urgent local correction is rushed through immediately to prevent error accumulation and redundant measurement point generation, while the comprehensive pose graph optimization is performed subsequently when time is less critical. This skipping approach prioritizes time-sensitive corrections while still completing the thorough optimization for overall accuracy.
Data Source
AI summary
An information processing apparatus performs first correction for correcting, in a case where a loop of a movement path of a sensor is detected, a position and orientation associated with a first measurement point used for estimation of a position and orientation to a position and orientation based on a second measurement point that is present near the sensor when the loop is detected.