Multi-Variate Attribute Value Frequency Algorithm for Identifying Anomalous Autonomous Emergency Braking Interventions
ARAI Journal of Mobility Technology · 16.07.2025 · 6 (3) · 2189-2196
DOI: 10.37285/ajmt.6.3.1 ↗Kurz gefasst
Fachartikel in der Zeitschrift ARAI Journal of Mobility Technology, online erschienen am 16.07.2025. Verfasst von Sriram Sridhar und 2 weiteren Autor:innen. Die Arbeit nennt 10 Literaturangaben.
Abstract (Originalsprache)
In general, Autonomous Emergency Braking (AEB) interventions globally are huge, as most vehicles are equipped with L2 or higher driving modes, making it impossible for the developers to evaluate these systems through manual analysis. This means that smart algorithms are essential to mine anomalous/interesting locations/locations with possibly faulty or unique system behavior by rarity of payload associated with the AEB events at those locations. Attribute Value Frequency (AVF) was originally proposed to solve this issue, as it identifies unique records provided that the features are categorical in nature. However, this algorithm, when used as is, is slow when applied to the context of big data, as is the case in our application to the ADAS domain. Also, in the original AVF, feature combinations are not considered, as each feature is considered in isolation, thereby diminishing its usability in our application. In this paper, we therefore propose the Multi-Variate Attribute Value Frequency (MVAVF) algorithm for data with categorical features / payloads, which not only takes feature pairs / combinations into consideration but also reduces computational complexity when applied to the Advanced Driver Assistance Systems (ADAS) domain, as unmatched feature combinations will be eliminated in the search space. Keywords: Autonomous Emergency Braking (AEB), ADAS , AVF algorithm, Multi-Variate, MV- AVF algorithm, Attribute Value Frequency (AVF), Multi-Variate Attribute Value Frequency (MVAVF)
Zitieren
APA
Sridhar, S., Podila, K. L., & Aswani, S. (2025). Multi-Variate Attribute Value Frequency Algorithm for Identifying Anomalous Autonomous Emergency Braking Interventions. ARAI Journal of Mobility Technology, 6(3), 2189-2196. https://doi.org/10.37285/ajmt.6.3.1
Vancouver
Sridhar S, Podila KL, Aswani S. Multi-Variate Attribute Value Frequency Algorithm for Identifying Anomalous Autonomous Emergency Braking Interventions. ARAI Journal of Mobility Technology. 2025;6(3):2189-2196. doi:10.37285/ajmt.6.3.1
BibTeX
@article{Sridhar2025,
author = {Sridhar, Sriram and Podila, Krishna Lalith and Aswani, Shelendra},
title = {Multi-Variate Attribute Value Frequency Algorithm for Identifying Anomalous Autonomous Emergency Braking Interventions},
journal = {ARAI Journal of Mobility Technology},
year = {2025},
volume = {6},
number = {3},
pages = {2189--2196},
doi = {10.37285/ajmt.6.3.1}
}
EUVIMED