AI-Based Car Fault Detection from Engine Sound
Supervisor Name
Ismail Khater
Supervisor Email
ikhater@birzeit.edu
University
Birzeit University
Research field
Machine Learning
Bio
Description
Drivers often cannot tell whether an unusual engine sound means a real fault, and a visit to a mechanic costs time and money. This project develops an AI-based system in which a driver records the car's engine with a mobile phone and the system classifies the sound as normal or as a specific fault, giving an early indication before a professional inspection. Building on the team's graduation project, the students will develop and compare deep learning models for engine sound, improve their robustness to real-world noise and different phones through data augmentation, enable the system to report when it is uncertain, and integrate the best model into a working mobile application.
