Map Image Feature Extraction for Autonomous Driving Traffic Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The high cost and time-consuming process of creating high precision maps for autonomous driving vehicles, which are prone to human errors, hinder effective traffic prediction and safe navigation.
Innovation Solution
A computer-implemented method using regular navigation maps with advanced vision technologies to identify features like roads, intersections, and roundabouts, allowing autonomous vehicles to predict traffic participant behaviors and plan trajectories.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If high precision maps are created using traditional methods, then map accuracy is improved, but cost and time consumption increase significantly
Solution Approach 1:
The patent uses satellite imagery and aerial photographs as copies or representations of the actual geographic environment to create map data. Instead of manually surveying and measuring physical locations, the system processes existing visual copies of the terrain to extract map features, significantly reducing time consumption while maintaining accuracy
Solution Approach 2:
The patent replaces manual mechanical labeling processes with automated image processing and machine learning algorithms. The system automatically identifies and extracts map features from images through computational methods, eliminating the need for human annotators to manually draw and label geographic features, thus reducing both time and labor costs
2Measurement precision
If high precision maps are created using traditional methods, then map accuracy is improved, but human errors increase
Solution Approach 1:
The system performs automated self-labeling of map features through machine learning models that independently identify and annotate geographic elements from images. The algorithm processes the data without human intervention, eliminating human errors associated with manual labeling while maintaining consistent and accurate map creation
Solution Approach 2:
The patent substitutes human manual work with automated computational systems. Machine learning algorithms and image processing techniques replace human annotators, eliminating the possibility of human errors in map creation while maintaining or improving accuracy through consistent automated processing
3Measurement precision
If high definition maps are used for traffic prediction, then prediction accuracy is improved, but system complexity increases
Solution Approach 1:
The patent extracts only the essential map features needed for traffic prediction from complete high-definition maps. Instead of using all detailed information from HD maps, the system selectively extracts relevant features such as road geometry, intersections, and traffic patterns, reducing data complexity while maintaining prediction accuracy
Solution Approach 2:
The patent divides the map processing task into separate modular components: image acquisition, feature extraction, map construction, and traffic prediction. Each module handles a specific aspect of the process independently, reducing overall system complexity by breaking down the complex HD map processing into manageable segments
Data Source
AI summary
In one embodiment, in response to perception data perceiving a driving environment surrounding an ADV, a map image of a map covering a location associated with the driving environment is obtained. An image recognition is performed on the map image to recognize one or more objects from the map image. An object may represent a particular road, a building structure (e.g., a parking lot, an intersection, or a roundabout). One or more features are extracted from the recognized objects, where the features may indicate or describe the traffic condition of the driving environment. Behaviors of one or more traffic participants perceived from the perception data are predicted based on the extracted features. A trajectory for controlling the ADV to navigate through the driving environment is planned based on the predicted behaviors of the traffic participants. A traffic participant can be a vehicle, a cyclist, or a pedestrian.