LIDAR Point Cloud Traffic Sign Detection for HD Map Updates
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional maps for autonomous vehicles lack precision and accuracy, and are often outdated due to the high cost and time required for their creation and maintenance, making it difficult for autonomous vehicles to safely navigate roads with changing conditions.
Innovation Solution
The development of high-definition (HD) maps that utilize LIDAR sensor data to detect traffic signs and other road features, combined with deep learning models for classification, allowing for real-time updates and precise navigation without relying on expensive survey teams.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional survey teams create maps using high resolution sensors, then map accuracy is improved, but the cost and time required increases significantly
Solution Approach 1:
The system enables autonomous vehicles to automatically detect, collect, and contribute map data during their normal operations. Vehicles self-update map information by detecting road signs, lanes, and other features using their onboard sensors, eliminating the need for dedicated survey teams to manually collect and update map data.
Solution Approach 2:
The system creates copies of map data from multiple autonomous vehicles and merges them to generate updated HD maps. Instead of creating maps from scratch using expensive survey equipment, the system replicates and consolidates data from numerous vehicle sensors to produce accurate, updated maps at minimal cost.
2Reliability
If survey fleets increase in number to capture road updates more frequently, then map freshness is improved, but the cost increases
Solution Approach 1:
Autonomous vehicles perform multiple functions: they transport passengers or goods while simultaneously collecting map data. The same vehicle fleet that provides transportation service also serves as the mapping system, eliminating the need for separate survey vehicles and enabling continuous map updates through normal operational activities.
Solution Approach 2:
The system leverages the existing operational fleet of autonomous vehicles to automatically update maps during their regular service. Vehicles continuously collect and transmit map data as they navigate roads, providing fresh map information without requiring additional dedicated survey vehicles or increasing fleet size.
3Area of stationary object
If GPS systems are used for location determination, then coverage area is improved, but accuracy deteriorates to over 100 meters
Solution Approach 1:
The system combines GPS data with LIDAR point cloud data and detected road features to determine vehicle location. By merging multiple data sources including global position information with precise local features like lane markings and road signs detected by LIDAR, the system achieves both wide coverage and high accuracy (within 30 cm or better).
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
Enables autonomous vehicles to navigate safely with high accuracy and up-to-date information, reducing latency and storage requirements while maintaining precise location and road feature detection, enhancing the reliability of autonomous driving systems.
Implementation Method 1
receiving a point cloud representing a region, the point cloud captured by a LIDAR sensor mounted on a vehicle
Data Source
AI summary
According to an aspect of an embodiment, operations may comprise receiving a point cloud representing a region. The operations may also comprise identifying a cluster of points in the point cloud having a higher intensity than points outside the cluster of points. The operations may also comprise determining a bounding box around the cluster of points. The operations may also comprise identifying a traffic sign within the bounding box. The operations may also comprise projecting the bounding box to coordinates of an image of the region captured by a camera. The operations may also comprise employing a deep learning model to classify a traffic sign type of the traffic sign in a portion of the image within the projected bounding box. The operations may also comprise storing information regarding the traffic sign and the traffic sign type in a high definition (HD) map of the region.