Page 48 - tech fest 2025 ב״ש חוברת תקצירים

48
Smart Traffic-Signal Optimization Using Reinforcement Learning
CS-C-17
Or Avital; oravital1@gmail.com Lotem Cohen; lotemcohen2000@gmail.com
Advisor: Dr. Dvir Ross
SCE - Shamoon College of Engineering, Be’er-Sheva
The goal of this project was to develop a simulation of traffic signals that adapt to real-time traffic conditions by using reinforcement learning algorithms. The system can minimize congestion by learning optimal signal-timing strategies through trial and error. By comparing the results with traditional traffic-control methods, the project contributes to research on intelligent transportation systems.
Keywords: adaptive signal timing, congestion reduction, intelligent transportation systems, machine learning in urban planning, real-time traffic management, reinforcement learning, simulation and modeling, traffic-signal optimization
Monitoring Fly Interactions
CS-C-18
Bar Rosenzweig; barro1@ac.sce.ac.il Elad Swisa; eladsw@ac.sce.ac.il
Advisor: Dr. Aviad Elyashar
SCE - Shamoon College of Engineering, Be’er-Sheva
Insect behavior plays a vital role in fields such as pest control, food production, medicine, and agricultural research. A modular framework was previously developed to analyze insect interactions, primarily used by researchers at the Hebrew University Faculty of Agriculture. This project focused on extending the capabilities of that framework in successfully detecting interactions, specifically relating to flies. We developed new modules for territory calculation within a defined area, which then detected territorial conflicts, extracted features from fly movement data, and trained a classifier to recognize fly behavior. Additionally, we built a Python software package with a desktop interface for broader accessibility and use.
Keywords: behavior classification, insect interaction, modular framework, network science, territorial analysis