Reference Driving Path Modeling for Roads Without Lane Markings
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for creating high-definition maps for vehicle navigation struggle with accuracy, especially in areas without clear lane markings, leading to unreliable and uncomfortable automated driving experiences.
Innovation Solution
A device that determines a reference driving path by analyzing multiple vehicle trajectories and interpolating points to identify typical driving behavior, allowing for accurate lane profiling and trajectory planning independent of road markings.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If HD maps are created using sensor-based recordings of lane boundaries from a fleet of vehicles, then map data can be obtained without complex surveying vehicles, but the accuracy and reliability of the determined target trajectory deteriorates in sections without sufficient lane boundaries
Solution Approach 1:
The patent introduces reference driving paths as an intermediary element that mediates between the available sensor data (lane boundaries) and the required target trajectory. These reference paths are pre-determined through clustering of historical driving data and serve as a reliable guide when direct sensor-based lane boundary detection fails or is insufficient, thus maintaining trajectory accuracy without requiring complex surveying vehicles
Solution Approach 2:
The patent performs preliminary determination of reference driving paths by clustering historical driving data before actual navigation occurs. This preliminary action creates a library of reliable reference paths that can be quickly retrieved and used during real-time operation, avoiding the need for complex real-time calculations and ensuring accurate trajectory determination even when lane boundaries are not visible
2Extent of automation
If target trajectory is determined by running centrally between lane markings from HD map data, then automated longitudinal and transverse guidance can be achieved, but the trajectory may not run on available lanes and appears uncomfortable to users
Solution Approach 1:
The patent applies local quality by determining multiple reference driving paths for different locations and conditions rather than using a single centralized path. Each reference path is specifically tailored to its local context based on historical driving behavior at that location, allowing the system to select the most appropriate reference path that matches natural driving patterns for that specific segment, thereby improving user comfort while maintaining automation
Solution Approach 2:
The patent changes the parameter of trajectory determination from geometric center calculation to statistical clustering of historical positions. By transforming the reference path from a purely geometric construct (center between markings) to a statistically-derived path (cluster center of actual driving positions), the system aligns automated guidance with natural human driving behavior, improving user comfort
3Reliability
If standard accuracy maps with 1-10 meter precision are used for route guidance, then basic navigation functionality is sufficient, but high-accuracy map data with less than 1 meter deviation is required for reliable automated driving control
Solution Approach 1:
The patent implements self-service by using the fleet of vehicles themselves to generate and continuously improve the HD map data. As vehicles navigate using the determined reference paths and trajectories, their actual driving data is fed back into the system to refine and update the reference driving paths. This self-improving mechanism allows the system to achieve high measurement precision through collective learning from multiple vehicles without requiring external surveying operations
Data Source
AI summary
A device for determining a reference travel path for a road section determines a plurality of measured travel paths of one or more vehicles for a corresponding plurality of passages through the road section, and subdivides a course of the road section into a sequence of support point levels. For each of the plurality of measured travel paths, the device determines a sequence of cut points of the respective measured travel path with the corresponding sequence of support point levels and, for each sequence of support point levels, determines a respective amount of support points based on the determined cut points with the respective support point levels. The device determines at least one reference travel path for the road section based on the sequence of sets of support points for the corresponding sequence of support point levels.