Enhance statistical features with changepoint detection for driver behaviour analysis

Jamal Maktoubian, Son N. Tran, Anna Shillabeer, Muhammad Bilal Amin, Lawrence Sambrooks

Research output: Book chapter/Published conference paperConference paperpeer-review

Abstract

Driver behaviour modelling is a critical field that addresses complex and dynamic driving behaviours on roads with the goal of enhancing road safety, reducing air pollution, and improving vehicle performance. Recent advancements in sensor technology and machine learning (ML) techniques have facilitated the capture and analysis of driver behaviour patterns. Nonetheless, the efficacy of ML models heavily relies on the quality of the data used. Therefore, developing feature extraction techniques that provide high-quality inputs is crucial. In this paper, we conceptualised, implemented, and evaluated a novel feature model called Changepoint-based Statistical Feature (C-bSF). Initially, we extracted various statistical functions from raw sensor data, which were then aggregated using lagging windows. Following this, a changepoint detection method was used to derive the C-bSF feature. We compared the performance metrics of this new approach with other feature extraction methods, demonstrating the superiority of C-bSF in driver behaviour classification tasks across three datasets.
Original languageEnglish
Title of host publicationPRICAI 2024: Trends in Artificial Intelligence
Subtitle of host publication21st Pacific Rim International Conference on Artificial Intelligence, PRICAI 2024 Kyoto, Japan, November 18–24, 2024
EditorsRafik Hadfi, Alok Sharma, Quan Bai, Patricia Anthony, Takayuki Ito
PublisherSpringer
Pages231-242
Number of pages12
Volume15281
ISBN (Electronic)9789819601165
ISBN (Print)9789819601158
DOIs
Publication statusPublished - 2025
Event21st Pacific Rim International Conference on Artificial Intelligence 2024: PRICAI2024 - Kyoto University, Kyoto, Japan
Duration: 18 Nov 202424 Nov 2024
https://www.pricai.org/2024/
https://www.dropbox.com/scl/fi/lreaqzj40nt7piqcfcuhr/PRICAI-2024-Whole-Program-16112024-Ver-5.pdf?rlkey=ga7dbh7us284ld3tybao8n66x&e=2&st=ru4ds4m6&dl=0 (Program)
https://www.pricai.org/2024/index.php/programs/accepted-papers (Proceedings)
https://link.springer.com/content/pdf/bfm:978-981-96-0116-5/1 (Front matter)

Publication series

NameLecture Notes in Artificial Intelligence
PublisherSpringer
Volume15281
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference21st Pacific Rim International Conference on Artificial Intelligence 2024
Country/TerritoryJapan
CityKyoto
Period18/11/2424/11/24
OtherThe Pacific Rim International Conference on Artificial Intelligence (PRICAI) is an annual international event which concentrates on AI theories, technologies and their applications in the areas of scientific, social, and economic importance for countries in the Pacific Rim. In the past, the conferences have been held in Nagoya (1990), Seoul (1992), Beijing (1994), Cairns (1996), Singapore (1998), Melbourne (2000), Tokyo (2002), Auckland (2004), Guilin (2006), Hanoi (2008), Daegu (2010), Kuching (2012), Gold Coast (2014), Phuket (2016), Nanjing (2018), Fiji (2019), Yokohama (2020, online), Hanoi (2021, online), Shanghai (2022, hybrid) and Jakarta (2023, hybrid).

PRICAI 2024 will be held in person in Kyoto, Japan. The Program Committee invites technical papers on substantial, original, and unpublished research in all aspects of Artificial Intelligence. PRICAI-2024 aims to bring together researchers, practitioners, educators and users in AI and related communities for in-depth intellectual exchanges, research cooperation and professional development.
Internet address

Fingerprint

Dive into the research topics of 'Enhance statistical features with changepoint detection for driver behaviour analysis'. Together they form a unique fingerprint.

Cite this