Road Edge Detection via Aerial Image and Telemetry Fusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current autonomous vehicle navigation systems lack precise geolocation of road edges, particularly on residential roads with minimal markings, relying on costly laser guidance or camera systems, and aggregate vehicle telemetry data, which limits broad-based operation.
Innovation Solution
A system and method that calculates a first likelihood of a roadway edge from an aerial image and a second likelihood from vehicle telemetry data, fusing these to identify a final likelihood, using probability distributions and 95th percentile analysis of data points from vehicles traveling in both directions, to enhance edge detection accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If laser guidance systems or on-board camera systems are used to identify road edges, then measurement precision is improved, but device complexity and cost increase
Solution Approach 1:
The patent uses vehicle telemetry data as an intermediary to infer road edge locations. Instead of directly detecting road edges with complex sensors, the system collects telemetry data from multiple vehicles and uses probability distributions to estimate road centerlines and edges, thereby avoiding the need for expensive laser guidance or camera systems while maintaining acceptable accuracy
Solution Approach 2:
The patent creates a virtual model of road geometry by aggregating and processing telemetry data from multiple vehicles. This digital representation of road edges and centerlines serves as a copy that replaces the need for physical sensing equipment, allowing the system to navigate without direct optical or laser detection of road markings
2Device complexity
If aggregate vehicle telemetry data is used to estimate road centerline, then device complexity is reduced, but measurement precision deteriorates
Solution Approach 1:
The patent merges telemetry data from multiple vehicles traveling in both directions to create a more accurate estimate of road centerline and edges. By combining data from opposing traffic flows and using probability distribution fusion, the system achieves higher measurement precision while maintaining low device complexity
Solution Approach 2:
The patent transforms raw telemetry data into probability distributions and uses statistical parameters (such as mean and standard deviation of vehicle positions) to estimate road geometry. This parameter transformation allows the system to extract meaningful geometric information from noisy aggregate data, improving measurement precision without requiring complex hardware
3Measurement precision
If vehicle telemetry data from multiple vehicles is aggregated, then measurement precision is improved, but loss of time increases
Solution Approach 1:
The patent pre-processes and stores telemetry data as probability distributions during normal vehicle operation. When road edge detection is needed, the system can quickly query pre-computed probability distributions rather than collecting and processing raw telemetry data in real-time, thereby reducing the time loss while maintaining high measurement precision
Solution Approach 2:
The patent uses a sufficient number of telemetry data points from a reasonable time window to achieve accurate road edge detection without collecting excessive data. By determining when enough data has been gathered to achieve the required precision threshold, the system avoids unnecessary data collection time while ensuring accurate results
Data Source
AI summary
A method to detect a roadway edge includes calculating a first likelihood of a roadway edge from an aerial image of a roadway by shifting a centerline of the roadway perpendicular to the centerline and overlapping the centerline with image gradients. A second likelihood of the roadway edge is determined using a vehicle telemetry fitting a probability distribution to telemetry points along the roadway. The first likelihood of the roadway edge and the second likelihood of the roadway edge are fused to identify a final likelihood of the roadway edge.