Smart Subway System
Enhancing metro systems with AI and real-time data analysis to improve management efficiency and passenger experience.
About Us
Ömer Alper Güzel
4th-year Computer Engineering student at TED University. Captain of ALECTED (Alternative Energy Car Team). Experience in Software Engineering, Computer Graphics, UI/UX, and Mobile App Development. He has worked in areas such as Parallel Programming and Game Programming during his studies. During his internships, he gained valuable experience in User Experience/Interface and Mobile Application Development. Additionally, he has been involved in design initiatives for the Ankara Metro and had the opportunity to observe both the Ankara and Istanbul metros.
Hüseyin Emre Çelik
3rd-year Computer Engineering student at TED University. Responsible for back-end development and server management of the university's application. He also works as a radio technician at the university. With a strong interest in trains and railways, Hüseyin Emre has contributed to maps and wiki entries about the Ankara Metro and has been actively engaged in related studies. His enthusiasm for railway systems brings valuable domain knowledge to the project.
Gökberk Eftal Ersoy
3rd-year Computer Engineering student at TED University. Focus on back-end development and cybersecurity. Board member of ALECTED (Alternative Energy Car Team of TED University). Gökberk is skilled in people management and has been involved in social media activities and engagements. He continues his contributions in these areas while developing his technical expertise in back-end systems and security protocols.
Muhammed Furkan Güneş
3rd-year Computer Engineering student at TED University. Skills in Java, Python, relational databases, and cloud computing. He has developed his technical expertise through his studies and various projects. He continues to work on improving his knowledge in these fields by applying what he has learned in practical projects, contributing significantly to database design and cloud infrastructure for the project.
Our Project
Comprehensive Metro Solution
SmartSubway aims to transform traditional metro systems into intelligent, efficient networks, enhancing both operational management and passenger experience. By leveraging real-time data analysis and artificial intelligence, specifically YOLO for object detection, our system provides actionable insights into crowd densities within trains and stations, enabling metro administrators to make data-driven decisions for optimizing resources and improving service reliability.
The system comprises two primary user-facing components: a Mobile Application for passengers and an Admin Panel for metro operators. These are supported by a robust backend infrastructure that includes AI-driven analytics, real-time data processing, and persistent data storage.
Admin Panel
The Admin Panel serves as a centralized command center for metro administrators, empowering them with comprehensive tools for efficient operational management. This web-based interface offers tools for operational control, including real-time monitoring of crowd densities, train locations, and station performance.
- Operational Monitoring with real-time data on train locations and passenger densities
- Dynamic Resource Management for schedule adjustments based on real-time demand
- Security and Safety Management with CCTV integration and alarm systems
- Energy Management through SCADA system integration and analytics
Mobile App
The Mobile Application is designed to be an indispensable companion for metro passengers. It offers real-time navigation with optimized route suggestions based on current conditions like crowd density and train schedules, making passengers' journeys easier and more efficient.
- Real-time navigation with crowd density-aware routing
- Accessibility features with AR navigation and voice guidance
- Real-time service updates and disruption notifications
- Integrated ticketing and payment functionality
How It Works
AI-Powered Analysis
Our system integrates two complementary AI technologies: YOLO (You Only Look Once) for visual analysis and Gemini-powered LLM for intelligent decision-making. YOLO processes video feeds to anonymously track passenger flow, while our Gemini LLM ingests this data alongside SCADA metrics to provide actionable recommendations for optimizing train schedules, energy usage, and passenger distribution in real-time.
Real-Time Data Processing
The system ingests data from various sources including CCTV cameras and SCADA systems. This information is processed and analyzed in real-time, providing accurate insights on train positions, passenger densities, and energy consumption patterns to both administrators and passengers.
Integrated Operations
The Admin Panel displays real-time data and analytics through interactive dashboards and maps, enabling operators to make informed decisions. Simultaneously, the Mobile App delivers personalized navigation guidance and service updates to passengers based on current conditions and individual journeys.
Unity Simulation Environment
Our Unity-based simulation environment allows us to test AI models in controlled conditions, generating realistic passenger flows and train operations without compromising privacy or operational security.
Train Simulation
3D model of metro trains with simulated passenger boarding/alighting behaviors and realistic movement physics.
Station Platform
Detailed station environment with simulated crowd dynamics and passenger flow patterns.
YOLO Person Detection
AI-powered detection of passengers in simulated station environment, achieving 66% accuracy in our testing.
Crowd Density Analysis
In future, heatmap visualization of passenger density across platform areas will be used for enabling optimization of passenger flow.
Mobile Application
Our mobile application provides passengers with real-time information and personalized guidance for a smoother metro journey.
Train Lines Screen
Train lines page with quick access to all lines.
Station Map
Interactive station maps with navigation guidance.
Train Schedule
Real-time departure and arrival information.
Crowd Indicator
Visual representation of current station crowding.
Presentation Video
SmartSubway project presentation for CMPE 492 lecture
YOLOv4 architecture used for real-time passenger detection
Technologies
YOLO Algorithm
Our system implements the YOLO (You Only Look Once) algorithm for real-time object detection, specifically to identify passengers in metro environments without compromising anonymity. This state-of-the-art algorithm processes video streams with minimal latency, enabling accurate crowd density analysis for both safety and efficiency optimization.
Data Integration & SCADA
Our system integrates with multiple data sources, including SCADA systems that provide real-time energy consumption metrics. This integration enables monitoring of traction power usage, station energy consumption, and overall system performance. Using RESTful APIs, we ensure seamless and secure data flow between the metro infrastructure and our applications.
React & React Native
Our Admin Panel is built with React.js, creating a responsive, component-based interface for metro administrators. The Mobile Application utilizes React Native for cross-platform development, ensuring a consistent experience across iOS and Android devices. Both interfaces integrate with backend services through optimized APIs to provide real-time updates and interactive features.
Unity 6 & Blender Simulation
We utilized Unity 6 as a real-time 3D development platform to create detailed simulations of metro stations and passenger flows. This allows us to test our AI models in controlled environments before deployment. Blender complemented this by enabling the creation of optimized 3D assets representing trains, platforms, and passenger models with high visual fidelity.
OpenStreetMap & Leaflet
Our mapping system leverages OpenStreetMap and Overpass Turbo to extract accurate geospatial data of subway lines and stations. Leaflet.js renders these into interactive maps in the Admin Panel, enabling real-time visualization of the metro network with custom overlays for operational monitoring and decision support.
Project Documents
Project Proposal
Project proposal for SmartSubway project with a focus on basic explanation of the project, its objectives, and the technologies to be used.
Project Specifications Report
PSR for this project with a focus on the project specifications, including the objectives, scope, and requirements of the project.
High-Level Design Report
HLDR for this project with a focus on high level design.
Project Analysis Report
PAR for this project with a focus on the analysis of the project such as the requirements, design, and implementation.
Low-Level Design Report
LLDR for this project with a focus on low level design.
Test Plan Report
TPR for this project with a focus on the test plan, including the testing strategy, test cases, and expected results.
Final Report - TEDU Deadline
This document outlines the objectives, scope, and methodology of our project. It serves as a comprehensive overview of our approach to enhancing metro systems through AI and real-time data analysis.
Final Report - IEEE SIU 2025
This document outlines the objectives, scope, and methodology of our project. It serves as a comprehensive overview of our approach to enhancing metro systems through AI and real-time data analysis.