Semantic Lane Segment Mapping for Occluded Autonomous Roads
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current autonomous driving technologies face challenges in accurately determining lane segments and associating vehicles or objects with them, especially in scenarios where parts of the road are occluded, leading to potential unsafe driving operations due to errors in 3D location estimation and failure to handle non-flat road surfaces.
Innovation Solution
A computer system in the autonomous vehicle processes images from cameras and map data to identify lane segments by subdividing the road into polygons, determining overlapping regions, and assigning identifiers, which allows for accurate rendering and filtering of important objects and vehicles, thereby improving computational efficiency and safety.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If 3D location estimation is used to determine lane segments, then autonomous driving navigation can be achieved, but accuracy deteriorates in occluded scenarios and on non-flat road surfaces
Solution Approach 1:
The patent divides the road into multiple lane segments by processing camera images to identify lane markings and boundaries. Each lane segment is defined by connecting boundary points along the lane path, creating discrete navigable units that improve localization accuracy compared to treating the entire road as a single entity.
Solution Approach 2:
The patent transitions from 2D image coordinates to 3D world coordinates by incorporating depth information and camera calibration data. This dimensional transformation allows accurate representation of lane segments on non-flat surfaces and resolves occlusions by understanding the spatial relationships between road features.
2Measurement precision
If detailed processing of all road features is performed, then accurate lane segment identification is achieved, but computational resources increase
Solution Approach 1:
The patent extracts only the essential features needed for lane segment identification from camera images, such as lane marking boundaries and key geometric points. By filtering out unnecessary visual information and focusing only on critical features, the system achieves accurate lane segmentation with reduced computational overhead.
Solution Approach 2:
The patent performs preliminary processing of camera images to pre-identify lane boundaries and key points before full lane segment construction. This preliminary extraction of essential geometric information reduces the complexity of subsequent processing steps and optimizes computational resource usage.
Data Source
AI summary
Techniques are described for autonomous driving operation that includes receiving, by a computer located in a vehicle, an image from a camera located on the vehicle while the vehicle is operating on a road, wherein the image includes a plurality of lanes of the road; for each of the plurality of lanes: obtaining, from a map database stored in the computer, a set of values that describe locations of boundaries of a lane; dividing the lane into a plurality of polygons; rendering the plurality of polygons onto the image; and determining identifiers of lane segments of the lane; determining one or more characteristics of a lane segment on which the vehicle is operating based on an identifier of the lane segment; and causing the vehicle to perform a driving related operation in response to the one or more characteristics of the lane segment on which the vehicle is operating.