Description

AI Real Time Public Bus Events System

Introduction
This project aims to design and implement a Real Time Public Bus Events Management system that uses AI-based video analysis, real-time monitoring, and modern web technologies to identify selected events occurring inside and around public buses. The system will analyze camera feeds locally, generate structured alerts, and transfer relevant information to a control center for immediate operational response. The solution will extend existing recording systems by adding real-time event detection instead of relying only on post-event investigation.
Background
Public transportation vehicles already use cameras, recording units, routers, and cellular communication for security and operational purposes. In many existing installations, the recorded video is mainly used for investigation after an event has occurred. This creates a gap between the moment an unusual event happens and the moment an operator becomes aware of it.
In this project, an AI-based event management system will be developed. The system will process video from cameras installed inside and outside the bus, detect selected events, and forward alerts, images, or short clips to a control center. The project will focus on a practical subset of the original event list, while the architecture will remain open for future expansion to additional events, voice analysis, and edge deployment on NVIDIA-based hardware.
Project Scope
• Hardware Integration
- Use existing HD/IP camera feeds or recorded bus video as input to the system.
- Prepare the system for local edge processing using NVIDIA Jetson Orin Nano or an alternative device.
- Prepare communication with the existing router/cellular infrastructure and a remote control center.
• Software Development
- Build a real-time frontend dashboard in React.js for control-center operators.
- Develop a Python FastAPI backend server to handle API routes, event logic and communication with the AI module.
- Store and manage event data, alert status and analysis results using MongoDB.
• AI and System Integration
- Use OpenCV for video processing and a YOLO-based model for object detection as the first AI layer.
- Develop an Event Engine that converts raw detections into structured bus events and alert levels.
- Use WebSocket communication for live event updates from the backend to the dashboard.
- Prepare the architecture for future Video & Voice analysis and custom trained models for complex events.
Student Requirements
• Basic programming knowledge or willingness to learn the required technologies from the beginning.
• Willingness to learn JavaScript, Python, React.js and FastAPI.
• Basic understanding of REST API design will be developed during the project.
• Teamwork.
• Independent learning.
• High motivation.
Development Tools
• Frontend: React.js, Vite, Axios, Tailwind / CSS - (for styling)
• Backend: Python, FastAPI, Uvicorn, WebSocket (for real-time updates), Pydantic
• RESTful API
• Database: MongoDB
• AI / Computer Vision: OpenCV, YOLO
• Edge Processing: NVIDIA Jetson Orin Nano or alternative device
• Additional Tools: Postman, GitHub, VSCode
Deliverables
• RT-PBEM Web App: A control-center dashboard for detected events and live status updates.
• Specification document.
• YouTube video (10 min)
• Poster
• Presentation
• GitHub link to the code
• Working prototype that demonstrates the complete flow from video input, through AI analysis and event generation, to an alert displayed in the control-center dashboard.
Name: Netanel Ben Hamo
Email: netanelbe@hit.ac.il
AI Real Time Public Bus Events System - Full Stack

AI and Machine Learning

Description

Emphasis in project execution

The project is has cooperation with the industry and combines meeting deadlines while being creative and focused on the task
