Automating Classification of Disengagements Using FoxGlove
Name
Pirjo Jõelo
Abstract
This thesis presents the Disengagement Classification Method (DCM) for analysing autonomous driving disengagements and its validation. The DCM aims to improve the current manual classification by leveraging expert knowledge and the functionality provided by the FoxGlove tool. This results in integrating a custom panel with FoxGlove Studio, which provides semi-automated disengagement categorisation and real-time data viewing. Compared to the manual method, this improves the efficiency of the analysis. Possible future developments involve improving the documentation for manual analysis and extending the classification of unplanned events to provide better automated analysis tools.
Graduation Thesis language
English
Graduation Thesis type
Bachelor - Computer Science
Supervisor(s)
Dietmar Alfred Paul Kurt Pfahl
Defence year
2024