Semi-Autonomous Driving Validation Using Intervention Hotspots
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Semiautonomous driving modes in vehicles often fail consistently at specific points due to incorrect sensor data, leading to vehicle deviation from the road, necessitating human intervention to correct errors and identify areas where highly automated systems are not suitable.
Innovation Solution
A method and device that analyze intervention data from multiple vehicles to identify positions where human interventions occur most frequently, allowing for the determination of correction parameters and improved driving modes, and classify interventions to assess error causes, enabling more reliable semiautonomous driving by correcting sensor data and determining approval zones for automated systems.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If semiautonomous driving mode is implemented, then automation level is improved, but reliability deteriorates due to consistent failures at specific points
Solution Approach 1:
The system collects intervention data from multiple vehicles and uses it to identify problematic positions, then feeds this information back to improve the semiautonomous driving mode through correction parameters or targeted data collection, creating a continuous improvement loop that enhances reliability while maintaining automation
Solution Approach 2:
The system processes multiple types of data (sensor data, position data, intervention data) through a unified analysis framework that identifies problematic positions and generates correction parameters applicable across the vehicle fleet, making the solution universally applicable to improve reliability
2Reliability
If intervention data from multiple vehicles is collected and analyzed, then reliability is improved by identifying error-prone areas, but device complexity increases
Solution Approach 1:
The system extracts only the essential elements needed for analysis (intervention data, position data, and relevant sensor data) from the complex vehicle operation data, focusing analysis on specific problematic positions rather than processing all possible data, thereby improving reliability without proportionally increasing complexity
Solution Approach 2:
The system segments the analysis by identifying specific positions where interventions occur and analyzing each position separately to determine correction parameters, breaking down the complex problem of improving overall reliability into manageable position-specific solutions
3Measurement precision
If sensor data is transmitted for analysis, then measurement precision is improved for identifying intervention positions, but loss of time increases due to data transmission and processing
Solution Approach 1:
The system performs preliminary filtering and selection of sensor data to include only relevant information for intervention analysis, and pre-identifies positions of interest before detailed analysis, reducing the time needed for data processing while maintaining measurement precision
Solution Approach 2:
The system transmits and analyzes only the partial set of data that is most critical for identifying intervention positions (position data and key sensor data at intervention moments) rather than complete continuous data streams, achieving sufficient measurement precision with reduced time loss
Data Source
AI summary
In a method for checking an at least semi-autonomous driving mode of a vehicle, intervention data are provided by a plurality of vehicles. The intervention data each represent a human intervention in an at least semi-autonomous driving mode of a vehicle. The intervention data includes sensor data at the time of the intervention and position data for a position at the time of the intervention. In accordance with the intervention data from the plurality of vehicles, a position is determined at which an increased number of vehicles registered a human intervention.