Road Wear Localization Using FMCW LiDAR and Neural Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional on-road localization methods for transportation vehicles face limitations in accuracy and reliability due to the quality and quantity of data processed, particularly in real-time applications, and struggle to distinguish between driving lanes and other road features like concrete dividers, especially when GPS signals are obstructed or lane markings are obscured.
Innovation Solution
The use of Frequency Modulated Continuous Wave (FMCW) LiDAR sensors in combination with deep neural networks to detect road wear reference lines, providing high-resolution measurements of road surface texture and enabling improved navigation by aligning vehicle wheel centers with detected road wear, even in conditions where lane markings are unclear or obscured.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional on-road localization methods use traditional sensors and GPS, then the system complexity is low, but the measurement precision and reliability deteriorate when GPS signals are obstructed or lane markings are obscured
Solution Approach 1:
The patent combines FMCW LiDAR technology with deep neural network processing to create an integrated road wear detection system. The LiDAR sensor captures high-resolution road surface data, which is then processed by neural networks to identify road wear reference lines and calculate lateral position, merging sensing and processing functions into a unified system that achieves high measurement precision without relying on GPS or traditional lane markings
Solution Approach 2:
The patent replaces traditional mechanical/optical lane marking detection systems with FMCW LiDAR-based road wear detection. Instead of relying on visual lane markings that can be obscured, the system uses electromagnetic wave-based LiDAR to detect subtle changes in road surface texture caused by wheel wear, substituting a more advanced physical principle to overcome the limitations of conventional approaches
2Reliability
If FMCW LiDAR sensors with deep neural networks are used to detect road wear reference lines, then the measurement precision and reliability improve, but the device complexity and computational requirements increase
Solution Approach 1:
The system uses the natural road wear patterns created by vehicle traffic itself as the reference framework for localization. The deep neural network automatically learns to identify these self-generated road wear reference lines without requiring external infrastructure or manual marking, allowing the road surface to serve its dual function of vehicle support and localization reference, thereby improving reliability without proportionally increasing system complexity
Solution Approach 2:
The deep neural network is pre-trained to recognize road wear patterns and reference lines before actual localization tasks. This preliminary training phase enables the system to quickly and reliably identify lateral position cues from road wear data during operation, reducing the computational burden during real-time localization and making the overall system more manageable despite the advanced technology involved
3Measurement precision
If high-resolution road surface texture measurements are taken in real-time, then the measurement precision improves, but the processing time and computational load increase
Solution Approach 1:
The deep neural network extracts only the most relevant features from the high-resolution LiDAR point cloud data, specifically focusing on identifying road wear reference lines and their characteristics. Instead of processing the entire high-dimensional point cloud, the system extracts essential lateral position information, maintaining measurement precision while significantly reducing the computational load and processing time required for real-time operation
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Enhances the robustness and reliability of on-road localization by providing accurate lateral position data, improving safety and navigation in complex road scenarios such as merge areas and roundabouts, and maintaining functionality even when traditional sensors fail or data becomes unavailable.
Implementation Method 1
The use of Frequency Modulated Continuous Wave (FMCW) LiDAR sensors in combination with deep neural networks to detect road wear reference lines, providing high-resolution measurements of road surface texture
Data Source
Figure 1A
Figure 1B
Figure 1C
AI summary
The invention relates to a technical improvement for providing localization for a transportation vehicle by detecting road wear reference lines in a roadway on which the transportation vehicle is travelling and controlling, guiding or otherwise facilitating alignment of the transportation vehicle wheel centers with the detected centers of the road wear.