SELECTIVE THREE-DIMENSIONAL LOCALIZATION AND NAVIGATION SYSTEMS AND METHODS

Number: 20250237519
Published: 2025-07-24
Filed: 2024-01-22
Assignee: DISNEY ENTERPRISES, INC.
Inventors: Milluzzi; Andrew Jesse et al.
CPC: G01C21/3638; G06T17/05
Verdict: Medium Hardware
Source: Google Patents · FreePatentsOnline

The keeper's note

3D localization/navigation system technique (robotics/ride-adjacent, Milluzzi HIL-sim cluster).

Abstract

A method may include receiving a 3D point cloud of a space, identifying points of the 3D point cloud at selective locations of the 3D point cloud, and comparing the points to a map of the space to localize an AGV within the space. A method may include receiving a 3D point cloud of a space, identifying multiple points of the 3D point cloud at respective beam angles from a sensor, and comparing the multiple points to a map of the space to localize an AGV within the space. A method may include receiving a 3D point cloud of a space, identifying first and second sets of points at respective first and second 2D planes, and comparing the sets of points to a map of the space to localize an AGV within the space. Additional methods and associated systems are also disclosed.

Background

FIELD

The present application relates generally to selective three-dimensional (3D) localization and navigation, and more specifically, for example, to selective 3D point cloud localization and navigation. BACKGROUND

When an Automated Guided Vehicle (AGV) is operating in a confined space and localizes from visible landmarks, other vehicles or objects in the space can occlude the AGV's ability to see such landmarks. This is especially true when a fleet of AGVs is operating in the same space, where the number of AGVs introduce multiple occlusions for a single AGV. Landmarks are typically contained on a two-dimensional (2D) plane, level with a scanning sensor, making occlusion easy. 2D scanning is typically utilized due to sensor availability and system complexity. Even vehicles equipped with three-dimensional (3D) scanners still tend to use a single 2D plane to localize and navigate due to computational complexity.

Regarding 3D scanning specifically, previous solutions would either try to match the entire 3D point cloud or match to a single 2D plane. Trying to match to an entire 3D point cloud has high computational overhead. Matching to a single 2D plane or level, or matching to only specific beacons or anchoring elements, suffers from occlusion issues, as noted above. BRIEF SUMMARY

According to various embodiments, a method includes receiving a 3D point cloud of a space, and identifying a point set of the 3D point cloud at selective locations of th

Claims

1. A method comprising: receiving a three-dimensional (3D) point cloud of a space; identifying a point set of the 3D point cloud at selective locations of the 3D point cloud; and comparing the point set to a map of the space to localize an automated guide vehicle (AGV) within the space, wherein the map is a map of known environmental features of the space. || 9. A method comprising: receiving a three-dimensional (3D) point cloud of a space; identifying first, second, and third points of the 3D point cloud at respective first, second, and third beam angles from a sensor; and comparing the first, second, and third points to a map of the space to localize an automated guide vehicle (AGV) within the space. || 15. A method comprising: receiving a three-dimensional (3D) point cloud of a space; identifying a first set of points at a first two-dimensional (2D) plane of the 3D point cloud; identifying a second set of points at a second 2D plane of the 3D point cloud; and comparing the first set of points and the second set of points to a map of the space to localize an automated guide vehicle (AGV) within the space. || 22. A method for localizing an automated guide vehicle (AGV) within a ride environment, the method comprising: capturing environmental data via a three-dimensional (3D) sensor; filtering the environmental data based on a two-dimensional (2D) localization; and comparing the 2D localization to known environmental information to determine a localization of the AGV.