Research and publications

PLACEHOLDER AAAAAAAAAAAAAAAAA

Master's thesis

Western University · Computer Science · 2026

Lightweight Static File Analysis for Pre-Installation Anomaly Screening in Automotive OTA Updates

Modern automotive update systems use signatures, authorization and integrity verification to establish delivery trust. My thesis investigates a separate question: whether the files inside an accepted package appear benign for the target software environment.
Data
Vendor Linux-based IVI files and Android-Automotive-derived update contents.
Features
File size, entropy, byte histograms, count features and byte 3-gram representations.
Models
Classical, generative and domain-adaptive anomaly-detection methods.
Evaluation
External malware, localized modifications, dispersed modifications and cross-domain testing.
View the thesis→

Thesis contributions

Research grounded in real update contents.

01

Real-data foundation

Evaluated the framework using actual vendor and platform-derived infotainment software contents rather than relying only on synthetic benign data.

02

Domain evaluation

Measured how anomaly-detection models behave within and across Linux-based IVI and Android Automotive software domains.

03

Limited-data adaptation

Investigated whether meta-learning could support adaptation to new software domains using small benign support sets.

04

Efficiency and scalability

Measured preprocessing costs and considered centralized and parallel deployment paths within an automotive OTA pipeline.

Publications

Published research

Peer-reviewed conference work

Conferences

Presentations and participation

2025

9th Cyber Security in Networking Conference

CSNet 2025 · Abu Dhabi, United Arab Emirates

Author and presenter

Presented research on using generative modelling to detect anomalous and potentially malicious files in automotive over-the-air update environments.

Research interests

Problems I want to continue exploring.

01

Automotive cybersecurity

Security challenges affecting connected vehicles, infotainment systems and over-the-air software updates.

02

Anomaly detection

Unsupervised and generative methods for identifying suspicious files when labelled domain-specific malware is limited.

03

Secure software delivery

Content-level screening that complements signatures, authorization, integrity verification and rollback protection.

04

Applied machine learning

Feature engineering, domain adaptation, experimental evaluation and practical deployment constraints.

Research discussion

Interested in automotive security or applied anomaly detection?

Get in touch→