Autonomous Vehicle Image Localization in GPS-Denied Environments
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles face challenges in determining their location when primary location services like GPS are unavailable due to signal interference or lack of coverage in environments such as forests or buildings, leading to potential navigation failures.
Innovation Solution
An incremental, iterative image comparison process is implemented using onboard sensors to capture and compare images of the vehicle's surroundings with stored images having known coordinates, expanding the geographic radius until a match is found or a maximum radius is reached, allowing for location determination without relying solely on GPS.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If GPS and other electromagnetic signal-based location services are used, then location determination is fast and accurate in open environments, but the system becomes unreliable in environments with signal interference or lack of coverage
Solution Approach 1:
The location determination system is segmented into multiple independent methods: electromagnetic signal-based location (GPS, cellular, Wi-Fi) and image-based location. Each method operates independently and can be used when the other is unavailable, improving reliability across different environmental conditions.
Solution Approach 2:
The autonomous vehicle is equipped with multiple location determination capabilities that serve universal purposes across different environments. The system can switch between GPS-based location and image-based location depending on environmental conditions, making the location system universally applicable whether in open spaces or signal-denied environments.
2Reliability
If image-based location determination is implemented as a backup system, then reliability in signal-denied environments is improved, but image processing requirements and computational complexity increase
Solution Approach 1:
Images are captured and stored in advance at known GPS coordinates during normal operation. When GPS becomes unavailable, the system only needs to compare current images against the pre-stored image database rather than processing images from scratch, significantly reducing computational complexity.
Solution Approach 2:
The system compares images within a limited geographic radius around the expected vehicle position rather than searching the entire image database. This partial action approach reduces processing requirements while still ensuring reliable location determination by focusing computational resources on the most likely location area.
3Measurement precision
If the geographic radius for image comparison is expanded to ensure location accuracy, then measurement precision is improved, but processing time and computational load increase
Solution Approach 1:
The system performs image comparison within a partial geographic radius around the expected vehicle position rather than searching the entire possible area. This limited search scope reduces processing time while maintaining sufficient location precision by focusing on the most probable location region.
Solution Approach 2:
The system uses GPS data and vehicle motion information to dynamically adjust the geographic radius for image comparison. When GPS is available, the radius is smaller; when GPS is unavailable, the radius is expanded just enough to ensure location accuracy is found, optimizing the balance between precision and processing time.
Data Source
AI summary
Systems, methods, and computer-readable storage media for determining autonomous vehicle location using incremental image analysis. An exemplary method can include identifying an expected position of an autonomous vehicle which is moving, and identifying, an actual position of the autonomous vehicle. The identifying of the actual position occurs by obtaining images of the autonomous vehicle's surroundings, initiating an iterative image comparison of those images to previously stored images within a given geographic radius of the autonomous vehicle, and iteratively extending the radius (and the pictures being compared) until a match is found or until the maximum radius is reached.