Road Marking Matching Before Curve Entry for Stable Vehicle Trajectories
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing autonomous driving systems face misrecognition issues with road markings when a vehicle approaches a curved road, leading to inappropriate target trajectory generation.
Innovation Solution
A determination device and method that restricts the comparison of recognized road markings to a range before a reference point, considering the switching point to the curved road, to prevent misrecognition and ensure accurate matching with map road markings.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If the host vehicle travels on a curved road or branch road, then the system can handle complex road geometries, but misrecognition is likely to occur for the road marking recognized from the camera image
Solution Approach 1:
The patent segments the road marking recognition process into different phases based on vehicle position: before entering curved/branch roads, during traversal, and after exit. By dividing the recognition range dynamically according to vehicle position and road geometry, the system maintains high accuracy in each segment while adapting to complex road configurations.
Solution Approach 2:
The system performs preliminary identification of curved roads and branch roads using map information before the vehicle actually enters them. This advance detection allows the system to prepare appropriate recognition strategies in advance, preventing misrecognition from occurring in the first place by adjusting the recognition range before the critical situation arises.
2Loss of information
If the system uses full recognition range for road marking comparison, then more information is available for trajectory generation, but misrecognition occurs at timing before host vehicle enters curved road
Solution Approach 1:
The patent implements dynamic adjustment of the road marking recognition range based on real-time vehicle position, road geometry, and curvature characteristics. The recognition range expands and contracts dynamically rather than remaining fixed, allowing the system to include sufficient information for reliable trajectory generation while excluding regions prone to misrecognition.
Solution Approach 2:
The system introduces map information as an intermediary layer between camera detection and trajectory generation. Map information provides prior knowledge about road geometry, curved roads, and branch roads, which mediates the comparison process by highlighting critical regions that require special attention, thus improving reliability without losing essential information.
3Measurement precision
If the system restricts recognition range to prevent misrecognition, then accuracy improves, but less information is available for trajectory generation
Solution Approach 1:
The patent applies different recognition strategies to different spatial regions: in critical regions (before and during curved/branch roads), the system uses restricted recognition ranges with map information guidance for high accuracy; in non-critical regions, the system can use broader recognition ranges to gather more information. This local differentiation optimizes both accuracy and information availability.
Data Source
AI summary
A determination device includes a storage medium that stores computer-readable instructions, and a processor connected to the storage medium, in which the processor executes the computer-readable instructions to recognize a road marking present in a moving direction of a vehicle, and determine whether the recognized road marking matches a map road marking based on map information stored in a storage unit, and when the vehicle approaches a curved road, the processor determines whether the recognized road marking matches the map road marking, with a restriction on a range of the recognized road marking to a range before a reference point taking into consideration a switching point to the curved road.