Sparse Map Construction for Autonomous Vehicle Navigation
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 increase storage and update demands.
Innovation Solution
The implementation of a system using cameras to analyze images and construct a sparse map for navigation, combining GPS data, sensor data, and image processing to identify road boundaries, wheels of target vehicles, and classify moving objects, allowing for efficient data management and navigation responses.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional mapping technology is used to navigate, then navigation capability is provided, but the volume of data needed to store and update the map becomes extremely large
Solution Approach 1:
The patent extracts only the essential navigational features from the environment (road boundaries, lane markings, traffic signs, pedestrians, vehicles) rather than storing complete high-definition maps. This selective extraction of critical information reduces data volume while maintaining navigation capability.
Solution Approach 2:
The patent segments the visual environment into discrete detectable objects and features (road edges, lane markings, traffic signals, pedestrians, vehicles) that can be independently identified and processed. This segmentation allows the system to process only relevant navigational elements rather than entire map datasets.
2Measurement precision
If vast volumes of data are collected and analyzed for autonomous navigation, then navigation accuracy is improved, but processing complexity and computational demands increase
Solution Approach 1:
The system extracts only critical navigational features from the visual data stream (road boundaries, lane markings, traffic signs, pedestrians, vehicles) rather than processing all collected data. This extraction approach maintains navigation accuracy by focusing on essential elements while reducing computational complexity.
Solution Approach 2:
The patent introduces an intermediary processing layer that translates complex sensor data into simplified navigational parameters and decisions. This intermediary layer processes visual information to extract meaningful navigational cues, reducing the complexity of downstream processing while maintaining accuracy.
3Loss of information
If complete map data is stored and updated continuously, then comprehensive navigation information is available, but storage requirements and update demands become daunting
Solution Approach 1:
The system extracts only essential navigational information from the environment (road boundaries, lane markings, traffic signs, pedestrians, vehicles) rather than storing complete map data. This extraction maintains navigation information completeness for critical elements while dramatically reducing storage requirements.
Solution Approach 2:
The autonomous vehicle performs real-time extraction and analysis of navigational features from its sensor data without relying on pre-stored comprehensive maps. This self-service approach allows the vehicle to generate its own navigational information from environmental observations, reducing storage and update demands.
Data Source
AI summary
The present disclosure relates to systems and methods for road edge detection and mapping, for vehicle wheel identification and navigation based thereon, and for classification of objects as moving or non-moving. Such systems and methods may include the use of trained systems, such as one or more neural networks. Further, autonomous vehicle systems may incorporate aspects of one or more of the disclosed systems and methods.