Local Navigation Alignment for Sparse Autonomous Driving Maps
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles face challenges in navigating efficiently due to the vast amount of data they need to process and store for navigation, including visual information, GPS data, sensor data, and map data, which can lead to limitations and safety concerns.
Innovation Solution
The system uses cameras to provide navigation features by analyzing images to construct and navigate with a crowdsourced sparse map, aligning navigation information from multiple vehicles, and fusing data from various sources to optimize navigation while ensuring safety and comfort constraints.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If autonomous vehicles use traditional mapping technology with detailed map data, then navigation accuracy is improved, but data storage requirements and processing complexity increase significantly
Solution Approach 1:
The patent divides the mapping system into multiple components: sparse map data stored centrally, local coordinate systems created for each road segment, and image-based alignment markers distributed throughout the environment. This segmentation allows navigation accuracy to be maintained through distributed information while reducing the need to store complete detailed maps.
Solution Approach 2:
The patent implements local coordinate systems for specific road segments rather than relying on a single global coordinate framework. Each road segment has its own coordinate system aligned using image markers, allowing navigation to be accurate locally while minimizing the amount of data that needs to be stored and processed globally.
2Reliability
If autonomous vehicles collect and process vast volumes of sensor data, GPS data, and map data, then navigation reliability is improved, but processing time and computational complexity increase
Solution Approach 1:
The patent pre-aligns navigation information from multiple vehicles into local coordinate systems during off-line processing, creating pre-processed sparse map data. During actual navigation, vehicles can quickly query this pre-processed data without needing to perform complex real-time alignment operations, significantly reducing processing time while maintaining reliability.
Solution Approach 2:
The patent introduces image markers and alignment features as intermediary elements between raw sensor data and navigation decisions. These markers serve as reference points that simplify the alignment and fusion of data from multiple sources, reducing computational complexity while improving the reliability of navigation data fusion.
3Reliability
If autonomous vehicles use comprehensive map data and sensor fusion, then safety is improved, but the complexity of the navigation system increases
Solution Approach 1:
The patent extracts only the essential navigation information needed for safe operation from comprehensive sensor data and map data. By creating sparse maps that contain only critical alignment markers and road segment information rather than complete environmental models, the system maintains safety through targeted data selection while reducing overall system complexity.
4Measurement precision
If autonomous vehicles store and update detailed maps continuously, then navigation accuracy is maintained, but data transmission and update overhead increase
Solution Approach 1:
The patent implements partial map updates by only transmitting and storing changes to sparse map data rather than complete map revisions. Vehicles update their navigation information with only the necessary local coordinate system adjustments and new alignment markers, maintaining navigation accuracy while significantly reducing data transmission overhead and energy consumption.
Data Source
Figure 1
Figure 2A
Figure 2B
AI summary
A computer-implemented method or a server for aligning navigation information from a plurality of vehicles is provided. It involves receiving navigation information from a first plurality of vehicles, wherein the navigation information from the first plurality of vehicles is associated with a common road segment, aligning the navigation information within a first coordinate system associated with the common road segment, wherein the first coordinate system is based on a first plurality of images captured by image sensors included on the first plurality of vehicles, storing the aligned navigational information in association with the common road segment, and distributing the aligned navigational information to one or more autonomous vehicles.