Page 89 - tech fest 2025 אשדוד חוברת תקצירים

Airborne Security System for Monitoring and Detecting Breaches
ME-A-09
Simon Turgeman; simonturgeman25@gmail.com Maria Kaman; mariakaman2@gmail.com
Advisors: Dr. Shayke Bilu1, Dr. Itzhak Levy1 1SCE - Shamoon College of Engineering, Ashdod
A novel, autonomous aerial surveillance platform has been developed for high-security zone monitoring. This lightweight system integrates with Pixhawk-based unmanned aerial vehicles to provide continuous perimeter observation capabilities. The primary detection mechanism is a specialized optical sensor coupled with machine learning algorithms optimized for aerial perspective analysis. Neural network architecture enables real-time identification of multiple threats, including unauthorized human intrusion, wildlife incursions, and physical barrier compromises. Upon threat detection, the system initiates a customizable alert protocol and maintains target tracking functionality. Key innovations include the miniaturization of hardware, optimization of power consumption, and development of robust classification algorithms, resistant to environmental variables. This approach significantly enhances security infrastructure while minimizing human monitoring requirements and constraints, compared to traditional fixed surveillance systems.
Keywords: AI, image recognition, intrusion detection, security system, surveillance drones
Hospital Assistance and Intelligent Navigation Robot
ME-A-10
Yonatan Almagor; yonatal@ac.sce.ac.il Ofir Yom Tov; ofiryo@ac.sce.ac.il
Advisors: Dr. Itzhak Levy1, Mr. Yogev Attias1 1SCE - Shamoon College of Engineering, Ashdod
Navigating in hospitals often presents significant challenges, leading to confusion and frustration for patients and visitors. To improve patient experience, we developed an autonomous robot designed for real-time assistance and intelligent guidance within dynamic hospital environments. Our system integrates LiDAR and a vision sensor for environmental mapping, obstacle avoidance, and human detection, alongside a differential drive mechanism for accurate maneuvering. The robot implements the RRT* algorithm, known for efficiently generating optimal paths in dynamic settings. Its overall performance, emphasizing obstacle avoidance, adaptive navigation, and human interaction, was evaluated in simulations conducted in “CoppeliaSim,” utilizing integrated physics engines. The results highlight the robot’s potential for significantly enhancing both the comfort and convenience of patients and visitors.
Keywords: autonomous robot, computer vision,“CoppeliaSim”, hospital navigation, LiDAR mapping, obstacle avoidance, path planning, RRT* algorithm
Book of Abstracts | 2025
89