Attributed Roadway Trajectories for Sensor-Independent Vehicle Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Self-driving road vehicles face limitations in extracting information from onboard sensors, particularly for data outside their field of view or that cannot be processed in real time, leading to incomplete navigation and collision avoidance capabilities.
Innovation Solution
The generation of attributed roadway trajectories, which are pre-processed sequences of three-dimensional coordinates with associated attribute values representing legal, physical, and navigational properties of the roadway, allowing for sensor-independent data use in vehicle control systems.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If onboard sensors are used to extract roadway information in real time, then the vehicle can detect current environmental conditions, but information outside the field of view or beyond processing capacity cannot be obtained
Solution Approach 1:
The patent pre-processes roadway data before the vehicle reaches those locations, creating attributed roadway trajectories in advance. This allows comprehensive roadway information to be prepared and stored beforehand, making it available when the vehicle needs it without requiring real-time sensor processing for every detail.
Solution Approach 2:
The patent introduces an intermediate data processing system that creates attributed roadway trajectories as a mediator between raw survey data and the vehicle's control system. This intermediary layer pre-extracts and organizes roadway information, reducing the burden on onboard sensors and real-time processing.
2Loss of information
If more sophisticated sensors and additional processing capability are used to overcome information extraction limits, then more roadway information can be obtained, but vehicle cost and complexity increase
Solution Approach 1:
The system performs data extraction and processing in advance during offline map creation, rather than requiring all processing to occur in real-time on the vehicle. This preliminary action transfers computational burden from the vehicle to the map creation system, reducing vehicle complexity.
Solution Approach 2:
The patent extracts roadway information attributes from raw survey data during the map creation process, separating the information extraction function from the vehicle's onboard systems. This extraction occurs beforehand, allowing the vehicle to use pre-packaged information rather than processing raw data in real-time.
3Loss of information
If real-time sensor processing is used to extract all roadway information, then complete navigation data can be obtained, but processing time and computational load increase beyond practical limits
Solution Approach 1:
The patent performs comprehensive roadway data processing and attribute extraction in advance during offline map creation. By doing this preliminary work before the vehicle needs the information, the system avoids the time constraint of real-time processing while ensuring all necessary navigation data is available.
Solution Approach 2:
The patent divides roadway information into discrete attributed trajectory points with specific attributes (speed limits, lane markings, curvature, etc.). This segmentation allows the information to be organized in manageable units that can be efficiently stored and retrieved without requiring processing of the entire roadway dataset in real-time.
Data Source
AI summary
An attributed roadway trajectory comprises at least one ordered series of attributed roadway trajectory points, which are spaced along a curve that is defined in a terrestrial coordinate frame and tracks an along-roadway physical feature of a real-world roadway segment portrayed in a point cloud. Any arbitrary attributed roadway trajectory point on the curve passes a predetermined proximity test for proximity to point cloud points discriminated as representing part of the along-roadway physical feature. Each attributed roadway trajectory point has at least one attribute value representing a characteristic of the real-world roadway segment at a position on the real-world roadway segment spatially associated with the attributed roadway trajectory point. A control system of a self-driving road vehicle can use the attribute value(s) to adjust control of the self-driving road vehicle based on sensor-independent roadway data (e.g. where the position of the attributed roadway trajectory point(s) remains outside of sensor range).