Vehicle Localization Using Map Exits and Lane Sensors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for determining vehicle localization on a road, particularly for advanced driver assistance systems (ADAS) and autonomous driving, face limitations due to the low accuracy of Global Navigation Satellite System (GNSS) positioning, leading to ambiguity in identifying the correct road type, which can restrict the availability of automated driving features.
Innovation Solution
A method that combines map data with lane and exit information, along with sensor data from vehicle sensors, to confidently determine vehicle localization on a road by ensuring the vehicle has not left a lane, thereby eliminating ambiguity and enhancing localization accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If GNSS positioning is used to determine vehicle localization, then the system is simple and easy to operate, but the positioning accuracy is low (within 50m) leading to ambiguity in road identification
Solution Approach 1:
The patent combines multiple data sources including GNSS positioning data, map data with lane and exit information, and sensor data from vehicle sensors to determine vehicle localization. This merging of multiple information sources resolves the ambiguity that arises from using GNSS alone, enabling accurate identification of the vehicle's specific road and lane location.
Solution Approach 2:
The patent introduces map data with detailed lane and exit information as an intermediary between the low-precision GNSS positioning and the required high-precision localization. By using the upcoming exit identification and lane information as intermediate steps, the system can determine whether the vehicle is on the correct road and lane even when GNSS accuracy is insufficient for direct localization.
2Measurement precision
If multiple data sources are combined to improve localization accuracy, then positioning precision is improved, but system complexity increases
Solution Approach 1:
The patent performs preliminary identification of upcoming exits from map data and positioning data before making the final localization determination. By pre-identifying the upcoming exit and determining whether lane leaving is required to exit, the system can use sensor data more effectively to confirm localization, thereby managing complexity through structured preprocessing of information.
Solution Approach 2:
The patent uses sensor data from the vehicle's sensor system as feedback to verify whether the vehicle has left its lane. This feedback mechanism allows the system to continuously monitor lane position and confirm localization accuracy, enabling the complex multi-source integration to produce reliable results through continuous verification.
3Reliability
If the system requires high confidence in localization to activate ADAS features, then reliability is improved, but the availability of features is reduced due to GNSS accuracy limitations
Solution Approach 1:
The patent merges multiple independent data sources (GNSS positioning, map data with lane/exit information, and sensor data) to achieve high confidence in localization. This combination allows the system to reliably determine whether the vehicle is on the correct road and lane, thereby enabling ADAS feature activation even in locations where GNSS alone would be insufficient.
Solution Approach 2:
The patent uses map data containing lane and exit information as an intermediary to bridge the gap between low-precision GNSS and the high-confidence localization required for ADAS features. By identifying upcoming exits and analyzing whether lane leaving is required, the system can achieve the necessary localization confidence to activate features on a broader range of road types.
Data Source
Figure 1a~1b
Figure 1c~1d
Figure 2a
AI summary
A method, a non-transitory computer-readable storage medium, a system and a vehicle for determining localization of a vehicle on a road are disclosed. Map data with respect to at least one road, one or more lanes of the at least one road, and one or more exits from the at least one road are received. Furthermore, localization data indicating that the vehicle is localized on the at least one road, and positioning data indicating a position of the vehicle on the at least one road are obtained. From the map data and the positioning data, an upcoming exit from the at least one road is identified. Based on the localization data, the map data and the identified exit, it is determined that for the vehicle to exit the at least one road at the identified exit, leaving of a lane by the vehicle is required. Sensor data are further received from a sensor system of the vehicle. On condition that the sensor data indicate that leaving of a lane by the vehicle can be excluded, it is determining that the vehicle is still localized on the at least one road.