Source-linked AI summary
On 'A Kalman Filter-Based Algorithm for IMU-Camera Calibration: Observability Analysis and Performance Evaluation'
Yuanxin Wu
TL;DR
The paper revisits a central observability result for IMU-camera calibration, arguing that its proof and resulting conclusion are incorrect. It corrects the sufficient conditions and examines a contradiction in the claimed observability result, while noting limits of instantaneous analysis.
Problem
The paper addresses an incorrect proof underlying a widely used local observability result for camera-IMU calibration.
Method
The paper revises Corollary 1 and Lemma 3 and checks the observability conclusion using Lie-derivative-based observability analysis.
Results
The corrected sufficient conditions require rotation with angular velocity having at least two nonzero components, but the claimed observability conclusion produces a contradiction in a three-component case.
Takeaways & Limitations
The paper challenges the original observability result and distinguishes its tighter rotation condition from the corrected condition, which generally needs one rotation.
Takeaways & Limitations
The analysis concerns instantaneous system properties and cannot represent rotations about different axes performed instantaneously; the authors also could not identify the source of the contradiction.
Abstract
from arXiv · showhide
The above-mentioned work [1] in IEEE-TR'08 presented an extended Kalman filter for calibrating the misalignment between a camera and an IMU. As one of the main contributions, the locally weakly observable analysis was carried out using Lie derivatives. The seminal paper [1] is undoubtedly the cornerstone of current observability work in SLAM and a number of real SLAM systems have been developed on the observability result of this paper, such as [2, 3]. However, the main observability result of this paper [1] is founded on an incorrect proof and actually cannot be acquired using the local observability technique therein, a fact that is apparently not noticed by the SLAM community over a number of years.