Sparse Road Mapping for Autonomous Vehicle Navigation Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles face challenges in navigating roadways due to the vast amounts of data required for processing and storing visual information, map data, and sensor data, which can limit their navigation capabilities and lead to inefficiencies in data storage and updating.
Innovation Solution
A method and system for autonomous vehicle navigation using a sparse map that includes polynomial representations of road segments and landmarks, allowing for efficient data storage and navigation without the need for extensive data storage or transfer, utilizing a combination of cameras, GPS, and sensors to generate and update the map through 'crowdsourcing' from multiple vehicle drives.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional mapping technology is used to store and update map data, then navigation accuracy is improved, but data storage requirements and bandwidth consumption increase significantly
Solution Approach 1:
The patent extracts only the essential geometric features of road segments (polynomial representations) and landmarks from complete map data, storing only these critical elements in the sparse map database while omitting redundant detailed information, thereby reducing data storage requirements while maintaining navigation accuracy
Solution Approach 2:
Instead of storing complete detailed maps and processing all that data, the patent inverts the approach by storing minimal sparse map data and generating detailed navigation information on-demand through polynomial representations and real-time sensor fusion, reducing stored data volume while preserving navigation capability
2Manufacturing precision
If complete map data is stored and updated frequently, then route accuracy is improved, but data transfer bandwidth and storage capacity requirements increase
Solution Approach 1:
The patent extracts only essential route geometry information (polynomial coefficients defining road segments) and key landmark positions from complete map data, storing minimal sparse representations that require far less bandwidth for transfer and storage while maintaining sufficient route accuracy for navigation
Solution Approach 2:
The patent changes the representation parameters of map data from detailed pixel-based or point-cloud formats to compact polynomial representations of road segments and simplified landmark geometries, reducing data size and bandwidth requirements while preserving essential route information
3Reliability
If vast volumes of sensor data and image data are processed, then navigation reliability is improved, but processing complexity and storage requirements increase
Solution Approach 1:
The patent extracts only relevant features from sensor data and images (such as lane markings, traffic signs, and obstacle positions) and integrates these with sparse map data, avoiding processing of all raw sensor data while maintaining navigation reliability through selective feature extraction and fusion
Data Source
Figure 1
Figure 2A
Figure 2B
AI summary
Systems and methods are provided for autonomous vehicle navigation. The systems and methods may map a lane mark, may map a directional arrow, selectively harvest road information based on data quality, map road segment free spaces, map traffic lights and determine traffic light relevancy, and map traffic lights and associated traffic light cycle times.