This topic can be chosen as an internship or a master thesis topic.

This thesis explores the use of recorded audio from train cars to extract meaningful insights into passenger behavior, movement patterns, seat occupation, and train functionality. By leveraging machine learning and signal processing techniques, the study aims to develop a system capable of analyzing audio data to detect key events such as passenger boarding and alighting, seat usage, crowd density, and mechanical anomalies.

The research will focus on feature extraction from audio signals, including sound classification, spatial localization, and anomaly detection. It will also address challenges such as noise filtering, privacy concerns, and the differentiation of relevant acoustic events from background sounds. The findings could contribute to improved passenger flow management, enhanced train maintenance, and more efficient public transportation operations.

Level: Bachelor, Master

Speciality: Machine Learning, Sound Analysis, Signal processing

Type of work: Research: 50%, Implemen: 30%, Experim: 20%

Type of activities: Experimenting, literature study, programming

Num of students: 1