Video-Based Vehicle Positioning With Local Map Alignment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current navigation and mapping systems for autonomous vehicles face challenges in accurately localizing vehicles within road networks, integrating data from diverse sensors, and maintaining up-to-date pre-built maps, which are often out-of-date and lack structural data for efficient map matching and object detection.
Innovation Solution
A method and system that utilizes camera images to generate a local map representation, align it with a reference map, and determine vehicle location and orientation, while updating the reference map with new data for improved accuracy and consistency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If pre-built maps are used for autonomous driving navigation, then global localization and route planning are enabled, but the maps become outdated and fail to reflect current road conditions
Solution Approach 1:
The system performs preliminary actions by continuously capturing road images and updating map data in advance before navigation needs occur. Road images are stored and processed proactively, so that when localization is needed, current map data is already available, preventing the map from becoming outdated.
Solution Approach 2:
The system establishes a feedback loop where ad-hoc sensor data from autonomous vehicles is continuously compared with pre-built maps, and discrepancies trigger automatic map updates. This closed-loop feedback ensures maps remain synchronized with actual road conditions, maintaining both reliability and up-to-dateness.
2Productivity
If ad-hoc sensing is used for environment perception, then real-time localization is achieved, but sensing errors accumulate and reduce navigation safety
Solution Approach 1:
The system uses pre-built maps as a reference feedback mechanism to verify and correct ad-hoc sensing data. By continuously comparing sensor-derived positions with map-based expectations, the system detects and corrects sensing errors, preventing error accumulation while maintaining real-time localization capability.
Solution Approach 2:
The pre-built map serves as an intermediary between ad-hoc sensors and the navigation system. It mediates by providing a trusted reference framework that validates sensor data, filtering out errors before they propagate to the motion planning system, thus enhancing navigation safety.
3Measurement precision
If high-resolution map data is stored for accurate localization, then matching precision improves, but memory consumption and computational load increase
Solution Approach 1:
The system segments the map into discrete road image tiles that can be independently stored and processed. Instead of loading entire high-resolution maps into memory, only relevant segments are retrieved and matched against sensor data, reducing memory consumption while maintaining localization precision through selective high-resolution processing.
Solution Approach 2:
The system applies partial action by processing only the portions of the map that are currently relevant for localization, rather than continuously processing the entire map. This selective approach reduces computational load while maintaining precision in the active region of interest.
Data Source
AI summary
A method and system for obtaining, from at least one camera associated with a vehicle traveling through a road network, a sequence of images of a road or a road environment, each image being associated with a location where that image is captured, generating a local map representation of an area of the road network using at least some images from the sequence of images and the locations associated therewith, the generating including: processing the at least some of the images to detect an object in the road or the road environment, determining at least one transformation for tracking the object between the at least some of the images and, based on the at least one transformation and the locations associated with the at least some of the images, generating a two- or three-dimensional representation of the object relative to the area of the road network, comparing the local map representation with a reference map covering the area of the road network, and determining, based on the comparison, a geographical location and an orientation of the vehicle within the road network.