Robust GPS-Vision Localization via Integrity-Driven Landmark Attention

Bhamidipati, Sriramya; Gao, Grace Xingxin

doi:10.48550/arxiv.2101.04836

Cited by 1 publication

(1 citation statement)

References 33 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The interpretable nature of the ICET algorithm and its enhanced noise modeling make it particularly well suited for safety-of-life applications such as automated driving. Developing a rigorous safety case usually requires the use of transparent algorithms, to identify and bound the risks of rare but potentially hazardous situations, which represent a risk to navigation integrity (Bhamidipati & Gao, 2022;Zhu et al, 2022). By comparison, methods that rely on machine learning, such as LO-Net (Q.…”

Section: Introductionmentioning

confidence: 99%

ICET Online Accuracy Characterization for Geometry-Based Laser Scan Matching

McDermott,

Rife

2024

navi

View full text Add to dashboard Cite

Distribution-to-distribution point cloud registration algorithms are fast and interpretable and perform well in unstructured environments. Unfortunately, existing strategies for predicting the solution error for these methods are overly optimistic, particularly in regions containing large or extended physical objects.In this paper, we introduce the iterative closest ellipsoidal transform (ICET), a novel three-dimensional (3D) lidar scan-matching algorithm that re-envisions the normal distributions transform (NDT) in order to provide robust accuracy prediction from first principles. Like NDT, ICET subdivides a lidar scan into voxels in order to analyze complex scenes by considering many smaller local point distributions; however, ICET assesses the voxel distribution to distinguish random noise from deterministic structure. ICET then uses a weighted least-squares formulation to incorporate this noise/structure distinction while computing a localization solution and predicting the solution-error covariance.To demonstrate the reasonableness of our accuracy predictions, we verify 3D ICET in three lidar tests involving real-world automotive data, high-fidelity simulated trajectories, and simulated corner-case scenes. For each test, ICET consistently performs scan matching with sub-centimeter accuracy. With this level of accuracy, combined with the fact that the algorithm is fully interpretable, this algorithm is well suited for safety-critical transportation applications. Code is available at https://github.com/mcdermatt/ICET.

show abstract

Section: Introductionmentioning

confidence: 99%