Lane Recognition Using 3D Road Structures
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing lane recognition systems fail to accurately detect the position of a subject vehicle within a lane when white lines are blurred or obscured, especially under adverse weather conditions or when the relationship between white lines and road structures changes.
Innovation Solution
A lane recognition apparatus and method that calculates the subject vehicle's position and estimates the lane and its position within the lane using three-dimensional road structures and two-dimensional patterns, incorporating a subject vehicle position calculation unit, a traveling lane estimation unit, and a final determination unit to provide accurate lane recognition.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If white lines are used for lane recognition, then the system works effectively under clear conditions, but the recognition accuracy deteriorates when white lines are blurred or obscured by weather conditions
Solution Approach 1:
The patent introduces road structures (side walls, guard rails, curbs) as intermediary reference objects to establish a relationship with white lines. By detecting these three-dimensional road structures and correlating them with lane information, the system can determine lane position even when white lines are blurred or obscured, thus maintaining recognition reliability and precision under adverse conditions
Solution Approach 2:
The patent transitions from two-dimensional white line detection to three-dimensional road structure detection. By using depth information from laser radars and surrounding environment detection apparatus to identify three-dimensional road structures, the system adds a spatial dimension to lane recognition, enabling accurate lane position determination independent of white line visibility
2Reliability
If only laser radar or surrounding environment apparatus is used for lane detection, then the system can detect road structures, but the position of the subject vehicle in the lane cannot be determined
Solution Approach 1:
The patent merges multiple information sources: white line detection results, three-dimensional road structure detection from laser radars, two-dimensional pattern recognition from cameras, and pre-stored road information from databases. By integrating these diverse data sources and correlating them through the subject vehicle position, the system recovers complete lane position information that would be lost if any single source were used alone
Solution Approach 2:
The system uses feedback loops to continuously refine lane position determination. Detection results from multiple sensors are fed back to update and correct the estimated subject vehicle position, which in turn improves subsequent lane recognition accuracy. This iterative feedback process ensures that lane position information is maintained even when individual detection sources are insufficient
3Measurement precision
If multiple detection methods are combined, then lane recognition accuracy improves, but the device complexity increases
Solution Approach 1:
The patent designs a multi-functional integrated detection system where a single control unit coordinates multiple sensors (laser radars, cameras, surrounding environment detection apparatus). This universal control architecture allows the same system to perform white line detection, road structure detection, pattern recognition, and lane position calculation, reducing overall system complexity despite the multiple detection functions
Solution Approach 2:
The patent segments the lane recognition task into distinct functional modules: white line detection module, road structure detection module, pattern recognition module, and position calculation module. Each module processes specific types of information independently, then results are integrated through the control unit. This segmentation allows complex detection functions to be managed through modular, independent components
Data Source
AI summary
A lane recognition apparatus according to the present invention includes: a subject vehicle position calculation unit that calculates a current position of a subject vehicle; a traveling lane estimation unit that estimates a lane in which the subject vehicle is traveling and a position, in a direction perpendicular to a traveling direction, of the subject vehicle in the lane based on the current position of the subject vehicle, subject vehicle surroundings information indicating a positional relationship between a road structure being a three-dimensional structure existing in relation to a road and the subject vehicle, and road information associating the road structure with a lane of a road in terms of a positional relationship therebetween; and a traveling lane final determination unit that determines the lane in which the subject vehicle is traveling and the position of the subject vehicle in the lane based on results of estimation.