Aerial View Generation for Autonomous Vehicle Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles face inaccuracies in generating aerial views of roads, limiting their ability to identify lanes, vehicles, and objects, especially at distances beyond 50 meters, which is problematic for vehicles requiring control actions like braking.
Innovation Solution
A method using a machine learning model to generate an aerial view from a forward-facing camera view, determining vehicle trajectory and road layout in a Frenet-Serret coordinate system, and identifying obstacles, enabling accurate lane detection and object placement up to 200 meters, crucial for autonomous truck braking and control.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If sensor data is used to generate aerial view from ego vehicle's perspective, then autonomous vehicle control is enabled, but accuracy deteriorates beyond 50 meters distance
Solution Approach 1:
The patent creates a virtual aerial view by copying and transforming the forward-facing camera image through geometric projection and coordinate system transformation. Instead of relying on inaccurate sensor fusion for aerial perspective, the system generates a synthetic aerial representation from the known forward view, maintaining accuracy across extended distances by using the accurate forward-facing camera data as the source.
2Measurement precision
If forward-facing camera view is used to generate aerial view, then measurement precision is improved, but device complexity increases due to machine learning model requirements
Solution Approach 1:
The patent replaces complex multi-sensor mechanical systems with a computational approach using a single forward-facing camera combined with machine learning models. Instead of physically capturing aerial views with multiple sensors, the system uses algorithmic transformation of forward-facing camera data, substituting mechanical complexity with computational processing.
3Adaptability or versatility
If aerial view range is extended to 200 meters for truck braking distance, then autonomous control capability is improved, but reliability deteriorates due to accuracy limitations
Solution Approach 1:
The patent performs preliminary geometric transformation and coordinate system conversion to generate the aerial view representation before it is needed for control decisions. By pre-processing the forward-facing camera data into an accurate aerial perspective using established geometric relationships, the system ensures reliability is maintained across the extended 200-meter braking distance required for autonomous truck control.
Data Source
AI summary
This application is directed to aerial view generation for vehicle control. A vehicle obtains a forward-facing view of a road captured by a front-facing camera of a vehicle and applies a machine learning model to process the forward-facing view to predict determine a trajectory of the vehicle and a road layout based on a Frenet-Serret coordinate system of the road for the vehicle. The trajectory of the vehicle is combined with the road layout to predict an aerial view of the road, and the aerial view of the road is used to at least partially autonomously drive the vehicle. In some embodiments, the machine learning model is applied to process the forward-facing view to determine a first location of each of an obstacle vehicle in the Frenet-Serret coordinate system. The first location of the obstacle vehicle is converted to a vehicle location on the aerial view of the road.