I'm a software engineering student and I've been diving into deep learning and automotive cybersecurity. I wanted to share a project I’ve been working on: CANomaly-LSTM.
Modern vehicles rely heavily on the CAN bus network, which inherently lacks built-in encryption or authentication. This makes it vulnerable to various spoofing and injection attacks. I built this project to experiment with using Long Short-Term Memory (LSTM) networks to learn the normal temporal patterns of CAN messages and flag anomalies.
The project is built entirely in Python. It focuses on sequential data processing to detect irregular traffic patterns that traditional rule-based systems or static firewalls might miss.
I'm continuously trying to improve my AI/ML skills, so I would really appreciate any feedback from the community on the code structure, the model architecture, or suggestions on better ways to handle and preprocess automotive time-series data.
Hi HN,
I'm a software engineering student and I've been diving into deep learning and automotive cybersecurity. I wanted to share a project I’ve been working on: CANomaly-LSTM.
Modern vehicles rely heavily on the CAN bus network, which inherently lacks built-in encryption or authentication. This makes it vulnerable to various spoofing and injection attacks. I built this project to experiment with using Long Short-Term Memory (LSTM) networks to learn the normal temporal patterns of CAN messages and flag anomalies.
The project is built entirely in Python. It focuses on sequential data processing to detect irregular traffic patterns that traditional rule-based systems or static firewalls might miss.
I'm continuously trying to improve my AI/ML skills, so I would really appreciate any feedback from the community on the code structure, the model architecture, or suggestions on better ways to handle and preprocess automotive time-series data.
Link: https://github.com/Yigtwxx/CANomaly-LSTM