United States Department of Transportation
i
Low-cost Real-Time Learning-based Localization for Autonomous Systems
-
2024-10-29
[PDF-4.67 MB]
Details:
-
Creators:
-
Corporate Creators:
-
Corporate Contributors:
-
Subject/TRT Terms:
-
Publication/ Report Number:
-
Resource Type:
-
Geographical Coverage:
-
Edition:Final Report (July 1, 2023-June 30, 2024)
-
Corporate Publisher:
-
Abstract:Robot localization is the problem of finding a robot’s pose using a map and sensor measurements, like LiDAR scans or camera images. It is crucial for any moving autonomous vehicle to interact with the physical world correctly. However, finding injective mappings between measurements and poses is difficult because sensor measurements from multiple distant poses can be similar. To solve this ambiguity, Monte Carlo Localization (MCL), the widely adopted method, uses random hypothesis sampling and sensor measurement updates to infer the pose. Other common approaches are to use Bayesian filtering or to find better-distinguishable global descriptors on the map. Recent developments in localization research usually propose better measurement models or feature extractors within these frameworks. On contrary, this project we propose a radically new approach to frame the localization problem as an ambiguous inverse problem and solve it with an invertible neural network (INN). We claim that INN is naturally suitable for the localization problem with many benefits, in terms of high accuracy (within 0.25m for city-scale maps), high-speed operation (>150Hz) and operates on low-cost embedded system hardware. We will demonstrate this on point-cloud and camera datasets with evaluation on indoor and outdoor localization benchmarks, and also deploy it on 1/10th scale and 1/2 scale autonomous vehicles to show real-time and scalable operation.
-
Format:
-
Funding:
-
Collection(s):
-
Main Document Checksum:
-
Download URL:
-
File Type:
