Lane-Level Data Extraction via Vehicle Dynamics Clustering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current vehicle navigation systems lack comprehensive and up-to-date lane-level data for road intersections and segments, relying on outdated mapping databases that are costly and time-consuming to update, which limits the effectiveness of automated and assisted driving systems.
Innovation Solution
The system employs GPS-generated vehicle dynamics data to derive lane-level information using clustering algorithms and probabilistic modeling, leveraging crowd-sourced data from multiple vehicles to provide real-time lane-specific turning rules and virtual trajectories, eliminating the need for dedicated survey vehicles and improving ADAS functionality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If dedicated survey vehicles are employed to update mapping database, then data accuracy is improved, but cost and time consumption increase significantly
Solution Approach 1:
The system enables ordinary vehicles to automatically contribute their GPS trajectory data to update the mapping database without requiring dedicated survey vehicles. Each vehicle serves itself and simultaneously contributes to the collective improvement of map data through automated data collection and submission processes.
Solution Approach 2:
Ordinary vehicles perform multiple functions: they serve their primary transportation purpose while simultaneously acting as mobile sensors for data collection. The system transforms standard vehicles into multi-functional units that both transport passengers and contribute to mapping database updates.
2Loss of information
If dedicated survey vehicles are used to collect roadway information, then data completeness is improved, but manufacturing cost increases
Solution Approach 1:
The system eliminates the need for expensive dedicated survey vehicles by enabling ordinary vehicles to automatically collect and submit roadway information. The infrastructure serves itself by utilizing the existing fleet of vehicles already present on the roads.
Solution Approach 2:
Instead of deploying specialized survey vehicles, the system creates virtual copies of data collection capabilities through software implementation in ordinary vehicles. The functional capability is replicated across the vehicle fleet rather than being concentrated in dedicated hardware platforms.
3Loss of information
If traditional mapping databases are used, then system complexity is reduced, but lane-level data detail is insufficient
Solution Approach 1:
The system segments roadway information into hierarchical levels: road-level navigation maps provide general guidance while lane-level separation data provides detailed turning window information. This segmentation allows the system to maintain simplicity at the navigation level while adding detail where needed for automated driving maneuvers.
Solution Approach 2:
The system adds a new dimension of lane-level detail to the traditional two-dimensional road map by incorporating vertical layering of information. Lane separation data and turning window parameters are superimposed on the base map structure, creating a multi-layered data architecture.
Data Source
AI summary
Presented are systems and methods for extracting lane-level information of designated road segments by mining vehicle dynamics data traces. A method for controlling operation of a motor vehicle includes: determining the vehicle's location; identifying a road segment corresponding to the vehicle's location; receiving road-level data associated with this road segment; determining a turning angle and centerline for the road segment; receiving vehicle data indicative of vehicle locations and dynamics for multiple vehicles travelling on the road segment; determining, from this vehicle data, trajectory data indicative of start points, end points, and centerline offset distances for these vehicles; identifying total driving lanes for the road segment by processing the trajectory data with a clustering algorithm given the turning angle and centerline; extracting virtual trajectories for the driving lanes; and commanding a vehicle subsystem to execute a control operation based on an extracted virtual trajectory for at least one driving lane.