Infrastructure Camera Positioning via Point Cloud Registration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current vehicle positioning systems are ineffective in GPS-denied environments, leading to inaccuracies and limitations in advanced driver-assist systems such as lane assist and automated valet parking.
Innovation Solution
A method and system that generate a vehicle-based point cloud of objects in proximity and an infrastructure-based point cloud from a sensor at a fixed location, registering these point clouds to determine the vehicle's position in a global coordinate system.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If GPS-based positioning is used, then positioning accuracy is maintained in open environments, but positioning fails completely in GPS-denied environments
Solution Approach 1:
The patent introduces infrastructure cameras as intermediary objects that mediate between the vehicle and the positioning system. These cameras capture images of the environment, which are then processed to extract visual features and determine vehicle position, serving as a bridge when GPS is unavailable
Solution Approach 2:
The patent replaces the GPS satellite-based electromagnetic signal system with a ground-based visual recognition system using cameras and image processing. This substitution enables positioning in GPS-denied environments by using visual features from infrastructure cameras instead of satellite signals
2Reliability
If infrastructure cameras are deployed, then positioning reliability improves in GPS-denied environments, but system complexity increases
Solution Approach 1:
The patent makes infrastructure cameras serve multiple functions: they capture images for visual feature extraction, provide geometric relationships for positioning, and can potentially serve other urban infrastructure purposes. This multi-functionality justifies the added complexity by maximizing utility
Solution Approach 2:
The patent creates a virtual copy of the physical environment through infrastructure camera images and processed visual features. This digital representation allows the vehicle to understand its position without direct GPS signals, reducing the need for complex alternative hardware systems
Data Source
AI summary
A method of determining the position of a vehicle includes generating a vehicle-based point cloud of objects in proximity to the vehicle, referenced to a vehicle-based coordinate system. The method also includes receiving an infrastructure-based point cloud, referenced to a global coordinate system, of objects detected by a camera mounted at a fixed location external to the vehicle, and registering the vehicle-based point cloud with the infrastructure-based point cloud to determine the vehicle position in the global coordinate system.