Lane Geometry Generation Using Map and Surround Vehicle Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing driver assistance systems struggle to accurately detect lanes in situations where camera-based information is unreliable, such as at intersections, on unpaved roads, or in rainy weather, leading to potential system malfunctions.
Innovation Solution
An in-vehicle electronic device that generates lane information by combining information on structures around the vehicle and driving information of surrounding vehicles, using sensors and prestored navigation data to create geometric information about the travel lane.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If lane detection relies solely on camera-based information, then the system is simple to implement, but it fails to accurately detect lanes in challenging conditions such as intersections, unpaved roads, or rainy weather
Solution Approach 1:
The patent merges multiple information sources (camera images, navigation map data, and surrounding vehicle position data) to generate lane information. By combining these diverse data sources, the system achieves reliable lane detection in challenging conditions where camera-only systems fail, while distributing the complexity across multiple existing components rather than concentrating it in a single detection mechanism.
Solution Approach 2:
The patent introduces an intermediary processing system that integrates navigation map information and surrounding vehicle data to supplement and verify camera-based lane detection. This intermediary layer acts as a mediator that provides additional contextual information, enabling the system to maintain reliability in conditions where direct camera detection is insufficient.
2Measurement precision
If the system uses multiple data sources (sensor information and navigation data) to generate lane information, then measurement precision improves, but device complexity increases
Solution Approach 1:
The patent makes existing system components multi-functional: the camera serves both primary detection and verification functions, navigation data serves both routing and lane position reference functions, and surrounding vehicle data serves both safety monitoring and lane inference functions. This multi-functionality improves measurement precision without proportionally increasing device complexity, as existing components perform multiple roles.
Solution Approach 2:
The system performs preliminary actions by pre-processing and integrating navigation map data and surrounding vehicle information before they are needed for lane detection. This preliminary preparation of data reduces the real-time processing burden, allowing the system to achieve high measurement precision without overwhelming complexity during critical detection moments.
3Adaptability or versatility
If the system operates in environments without clear lane markings, then adaptability improves, but measurement precision deteriorates
Solution Approach 1:
The patent transitions from two-dimensional camera image analysis to three-dimensional spatial reasoning by incorporating navigation map data (which provides overhead, map-level perspective) and surrounding vehicle position data. This dimensional shift allows the system to infer lane boundaries in environments where traditional 2D visual cues are absent, achieving both environmental adaptability and maintained measurement precision through multi-dimensional data fusion.
Data Source
AI summary
The present disclosure provides a device and method for generating lane information. The method includes obtaining information on a structure located around a vehicle and driving information of a surround vehicle based on information collected by a sensor and prestored navigation information; and generating geometric information on a travel lane of the vehicle in a road area based on the information on the structure and the driving information of the surround vehicle.