Aerial Telemetry Fusion for Autonomous Map Generation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for generating high definition maps for autonomous vehicles are time-consuming, costly, and limited to public roads, as they require specialized driving mapping vehicles, which are not feasible for private roads and are inefficient in data collection and maintenance.
Innovation Solution
The method involves data fusion of aerial image data and telemetry data to generate map features such as lane width, curvature, and lane markings, using a processing system that includes a feature extraction and data processing engine to control vehicles, without the need for specialized mapping vehicles, enabling scalable and cost-effective map creation and updates.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If specialized driving mapping vehicles are used to collect data for high definition maps, then map accuracy and completeness are improved, but cost and time consumption increase significantly
Solution Approach 1:
The patent uses aerial images as a copy or alternative representation of the road environment, processing these images to extract map features without requiring physical presence of mapping vehicles on the road. This copying approach maintains map accuracy while significantly reducing time consumption.
Solution Approach 2:
The patent replaces the mechanical system of driving mapping vehicles physically traveling on roads with an aerial imaging and image processing system. This substitution eliminates the time-consuming aspect of ground-based data collection while maintaining the ability to extract accurate map features through computational methods.
2Measurement precision
If specialized driving mapping vehicles are deployed, then high definition map data is obtained, but cost increases and scalability to private roads is limited
Solution Approach 1:
The aerial imaging system serves multiple purposes: it can map both public and private roads, and the same system can be used for various types of map feature extraction. This universality increases cost-effectiveness by eliminating the need for specialized ground vehicles for each mapping project.
Solution Approach 2:
The patent employs standard aerial imaging technology rather than expensive specialized mapping vehicles. The approach uses readily available aerial images and processes them through computational methods, significantly reducing the cost barrier for map creation while maintaining adequate quality for autonomous navigation.
3Reliability
If ground survey data is used to update maps, then map accuracy is maintained, but the process requires physical mapping vehicle presence and becomes inefficient
Solution Approach 1:
For map updates, the patent uses aerial images as a copy of the current road state, processing these images to detect changes and update map features. This approach maintains map accuracy by using fresh aerial data while dramatically improving update efficiency by eliminating the need for ground survey vehicles.
4Productivity
If aerial image data and telemetry data are fused, then map feature generation becomes cost-effective and scalable, but data processing complexity increases
Solution Approach 1:
The patent introduces an intermediary processing layer that fuses aerial image data with telemetry data. This intermediary system reconciles the two data sources, using the aerial images as the primary map feature source and telemetry data to provide contextual information such as vehicle position and road geometry, thereby managing processing complexity while achieving scalable map creation.
Data Source
AI summary
In one example implementation according to aspects of the present disclosure, a computer-implemented method for generating map features includes receiving, by a processing device, aerial image data. The method further includes receiving, by the processing device, telemetry data. The method further includes performing, by the processing device, data fusion on the aerial image data and the telemetry data to generate map features. The method further includes controlling, by the processing system, a vehicle based at least in part on the map features.