Lane-Level Navigation Mapping Using SD Maps and Vehicle Pose
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional road-level navigation based on standard definition (SD) maps cannot provide lane-level guidance, and high-precision (HD) maps are costly and require high-precision sensors and computing power, limiting the availability of lane-level navigation services.
Innovation Solution
A method to construct lane-level navigation maps using standard definition maps by mapping road vectors and lane elements into a vehicle's body coordinate system, utilizing vehicle pose data and inverse perspective mapping, without requiring HD maps or high-precision sensors, using a target relational expression to enhance frame estimation accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If high-precision (HD) maps are used to provide lane-level navigation services, then navigation precision is improved, but system cost increases
Solution Approach 1:
The patent segments the lane-level navigation problem into two parts: using SD maps for general road-level navigation and overlaying detected lane elements (lane lines, road edges) for lane-level guidance. This segmentation allows the system to achieve lane-level precision without the high cost of complete HD maps, resolving the contradiction between navigation precision and system cost.
Solution Approach 2:
The patent creates a virtual lane-level map by copying and transforming detected lane elements from the real world into the map coordinate system. This virtual overlay provides lane-level navigation information without requiring expensive HD map data, thus improving navigation precision while controlling system cost.
2Measurement precision
If high-precision sensors and strong computing power are used to achieve lane-level positioning, then positioning accuracy is improved, but device complexity increases
Solution Approach 1:
The patent enables the navigation system to self-generate lane-level information by detecting lane elements using the vehicle's existing camera and processing them through inverse perspective mapping. This self-service approach eliminates the need for expensive external HD maps and high-precision sensors, achieving positioning accuracy while reducing device complexity.
Solution Approach 2:
The patent replaces the mechanical/sensor-based HD map system with a computational approach using inverse perspective mapping and coordinate transformation. This substitution uses software processing of standard camera images instead of expensive hardware sensors, maintaining positioning accuracy while reducing device complexity.
3Ease of manufacture
If standard definition maps are used for navigation, then system cost is reduced, but navigation precision deteriorates
Solution Approach 1:
The patent merges SD map data with real-time detected lane elements (lane lines, road edges) to create a composite navigation display. This merging combines the cost advantage of SD maps with the precision benefit of lane-level detection, achieving both low system cost and high navigation precision simultaneously.
Solution Approach 2:
The patent adds a lane-level dimension to the traditional road-level SD map navigation by overlaying detected lane elements. This dimensional enhancement transforms the navigation from generic road-level guidance to precise lane-level guidance while maintaining the use of inexpensive SD maps, thus improving precision without increasing cost.
Data Source
AI summary
A navigation method includes: obtaining a driving image and a pose of a vehicle at a first moment, and a standard definition map used for navigation; mapping a target element and a vector direction of a road on the standard definition map into a vehicle body coordinate system of the vehicle, where the target element includes a lane line or a road edge in the driving image; and constructing a lane-level navigation map at the first moment based on the vector direction of the road, the target element, the pose of the vehicle at the first moment, and a pose of the vehicle at a second moment, where the second moment is earlier than the first moment, and the lane-level navigation map is used to provide a lane-level navigation service for a user.