Vehicle Localization via Semantic Map Matching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing vehicle localization methods in autonomous driving are affected by changes in light intensity and object appearance, leading to poor accuracy and robustness, especially when GPS signals are unavailable.
Innovation Solution
A vehicle localization method that acquires images of the scene, determines semantic elements, matches them with map elements using category-specific methods, and calculates localization information using a deep learning model and sensor fusion algorithms.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If existing vehicle localization methods are used, then the system can operate without GPS signals, but the localization accuracy deteriorates due to changes in light intensity and object appearance
Solution Approach 1:
The patent segments the localization task by identifying and matching multiple independent semantic elements (landmarks, road markings, signs) in the scene with corresponding map elements. Each semantic element is processed and matched separately, then combined to determine overall vehicle position, thereby improving accuracy while maintaining reliability without GPS
Solution Approach 2:
The patent changes the parameter of feature representation by using deep learning models to extract robust semantic features that are invariant to changes in light intensity and object appearance. This allows the system to maintain high localization accuracy despite environmental variations, while still operating reliably without GPS signals
2Measurement precision
If semantic element matching with map elements is implemented, then localization accuracy improves, but the system complexity increases due to multiple processing steps
Solution Approach 1:
The patent implements a universal deep learning-based semantic segmentation model that handles multiple types of elements (landmarks, road markings, signs) through a single framework. This multi-functional approach improves localization accuracy by comprehensively matching various scene elements with map features, while avoiding the need for separate specialized processors for each element type, thus controlling system complexity
Data Source
AI summary
A vehicle localization method and device, an electronic device and a storage medium. The vehicle localization method includes: acquiring an image of a scene where a vehicle is located, the image comprising a semantic element; determining an element category corresponding to the semantic element; matching the semantic element with a map element to acquire a matching result by a matching method corresponding to the element category; and determining localization information of the vehicle according to the matching result.