Autonomous Vehicle Route Coverage Evaluation for Training Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for generating training data sets for autonomous vehicles face challenges in efficiently evaluating test runs due to deviations from planned routes and unforeseen events, requiring significant planning effort and additional evaluation steps, especially when using a fleet of test vehicles.
Innovation Solution
A method utilizing telemetric data to determine test run and target route tolerance ranges through interpolation, allowing for automated evaluation of route coverage and adaptive scheduling of additional test runs, reducing computational effort and improving data set quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If test runs are conducted with a fleet of test vehicles to obtain comprehensive training data, then the quality and quantity of training data sets improve, but the planning effort and evaluation complexity increase significantly
Solution Approach 1:
The system automatically evaluates test run data by comparing actual vehicle positions with scheduled route points, eliminating the need for manual evaluation. The automated process independently determines route coverage values and identifies deviations, allowing the system to self-assess test run quality without external intervention
Solution Approach 2:
The system establishes a feedback loop where test run results are automatically compared against scheduled routes, route coverage values are calculated, and deviations are identified. This feedback mechanism enables continuous improvement of test run planning by providing quantitative data on route adherence and data collection effectiveness
2Measurement precision
If manual evaluation steps are added to check whether driving specifications are met, then the measurement precision of test run assessment improves, but the time and computational resources required increase
Solution Approach 1:
The patent replaces manual evaluation processes with automated computational methods. Instead of human analysts reviewing test run data, the system uses automated algorithms to compare actual vehicle positions with scheduled routes, calculate route coverage values, and assess whether driving specifications are met, significantly reducing evaluation time while maintaining precision
Solution Approach 2:
The system transforms qualitative assessment criteria into quantitative parameters by calculating route coverage values based on actual vehicle positions and scheduled route points. This parameter transformation enables precise, automated evaluation of test run quality without requiring manual interpretation
3Extent of automation
If telemetric data is collected and processed to determine route coverage, then the automation extent of test run evaluation improves, but the computational effort required increases
Solution Approach 1:
The system extracts only the essential telemetric data needed for route coverage evaluation - specifically actual vehicle positions - from the complete set of test run data. By focusing only on position information rather than processing all available sensor data, the system achieves effective automation while minimizing computational energy consumption
Data Source
AI summary
Training data sets for an autonomous vehicle are generated using a test run performed by a test vehicle along a measured route during which vehicle environment data is detected along the measured route by the test vehicle and telemetric data, including location coordinates of several actual vehicle positions reached by the test vehicle, of the test vehicle is time-discretely transmitted to an external telemetric data collection unit during the first test run. A test run tolerance range or a target route tolerance range is determined. A route coverage value is determined and if the route coverage value is lower than a threshold, an updated scheduled route is created for a second test run with the first test vehicle or a further test vehicle.