Visual-Inertial Navigation Reference Frame Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Visual-inertial navigation systems face challenges in maintaining robustness, particularly when vision constraints are lost due to fast motion, camera shielding, or lack of texture, leading to errors in state vector calculations and trajectory drift.
Innovation Solution
The method involves determining a reference frame, correcting it by introducing a new vision constraint when the system enters a stationary state, and performing non-linear optimization for frames other than the reference frame and the oldest frame to prevent pose drift.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the visual-inertial navigation system relies solely on IMU constraint when vision constraint is lost, then the system can continue to operate, but the calculated state vectors are subject to accumulated errors and trajectory drift occurs
Solution Approach 1:
The system performs preliminary actions by identifying stationary states in advance and introducing new vision constraints before significant drift occurs. The method detects when the system enters a stationary state and proactively corrects the reference frame using vision information, preventing accumulated errors from degrading measurement precision.
Solution Approach 2:
The system implements feedback by continuously monitoring the validity of vision constraints and the stationary state of the system. When vision constraints are lost, the system detects this condition and feeds back correction mechanisms by introducing new vision constraints based on stationary frame relationships, thereby maintaining state vector accuracy.
2Measurement precision
If the system performs non-linear optimization for all frames, then the computation is thorough, but the computation time increases and real-time performance deteriorates
Solution Approach 1:
The system extracts and excludes the reference frame and oldest frame from non-linear optimization when the system is in a stationary state. By taking out these frames that do not require optimization, the system reduces computation time while maintaining optimization accuracy for the remaining frames that are critical for navigation.
Solution Approach 2:
The system applies partial optimization by performing non-linear optimization only on necessary frames (excluding reference frame and oldest frame) rather than all frames. This partial action approach maintains sufficient optimization accuracy for real-time performance by focusing computational resources on frames that most impact navigation accuracy.
Data Source
AI summary
A method for improving robustness of a visual-inertial navigation system includes: determining a reference frame; correcting the reference frame; and performing non-linear optimization for frames other than the reference frame and an oldest frame according to the oldest frame and the corrected reference frame.