Skip to content

Pranjulchaurasiya/Buildathon-Project

Repository files navigation

AI-Powered Decision-Support System for Train Traffic Control

Overview

This project was developed for the Gemini & Firebase Buildathon organized by TensorFlow User Group Ghaziabad.
It is an AI + Optimization powered intelligent decision-support system that assists railway section controllers in making optimized, real-time decisions for train precedence and crossings.


Problem Statement

Indian Railways currently relies on manual decision-making by section controllers. With rising traffic, limited track capacity, and frequent disruptions (delays, breakdowns, weather), this manual system struggles to maintain punctuality, safety, and throughput.

Challenges:

  • Manual precedence decisions under pressure
  • Limited infrastructure shared by express, passenger, and freight trains
  • Real-time disruptions affecting efficiency
  • Lack of intelligent decision-support tools

Proposed Solution

An intelligent decision-support system that combines AI + Operations Research + Firebase Cloud tools to:

  • Generate conflict-free, feasible train schedules
  • Re-optimize in real time under delays or disruptions
  • Provide What-If simulations powered by Gemini AI
  • Deliver recommendations (Proceed / Hold / Reroute) via a dashboard
  • Track performance KPIs like punctuality, average delay, and throughput

Core Features

  1. Optimization Engine (Google OR-Tools) → Generates conflict-free train schedules.
  2. What-If Simulation (Gemini AI) → Natural language disruption analysis.
  3. User-Friendly Dashboard (React/Streamlit + Firebase Hosting) → Visual train timelines, controller recommendations, overrides.
  4. Realtime Data Sync (Firebase Firestore) → Instant updates across dashboards.
  5. Performance KPIs (Looker Studio) → Live dashboards for delay, throughput, and utilization.

Technology Stack

  • Gemini AI → Natural language simulation & explanations
  • Firebase Firestore → Realtime database & audit logs
  • Firebase Hosting → Dashboard deployment
  • Firebase Auth → Secure user login
  • Firebase Functions → API backend for optimization engine
  • Google OR-Tools → Integer Linear Programming for scheduling
  • Looker Studio → Live KPI dashboards
  • React.js / Streamlit → Frontend interface

Architecture Overview

Screenshot 2025-09-09 145949

Implementation Roadmap

  • Phase 1 (Hackathon): Section-level prototype with CSV inputs, OR-Tools optimization, Firebase dashboard
  • Phase 2: Real-time delays + Gemini-powered What-If simulation
  • Phase 3: Multi-section optimization + predictive analytics
  • Phase 4: Nationwide integration with Indian Railways TMS

Innovation & Uniqueness

  • Human-in-the-loop design: AI assists, controller decides
  • AI-powered What-If analysis with Gemini
  • Google ecosystem-only build: scalable, cloud-native
  • KPI tracking + audit logs for accountability

Expected Impact

  • Economic: Reduced fuel use, lower operational costs, efficient freight
  • Environmental: Lower emissions from reduced idle time
  • Operational: Stress-free controllers, improved punctuality
  • Scalable: Section-level prototype → national adoption
  • Community: Reliable railways, improved public trust

References


License

This project was created as part of the Gemini & Firebase Buildathon. For research and educational purposes only.