🚆 AI Co-Pilot for Indian Railways

Smart India Hackathon 2025 – Team Falcon
Problem Statement ID: SIH25022 | Theme: Transportation & Logistics | Category: Software

---

🌍 Vision

Our solution is an AI-powered Decision Support System designed for the unique challenges of Indian Railways.
It acts as an intelligent co-pilot for traffic controllers, upgrading the manual system into a modern, proactive one.

👉 The goal is not to replace the invaluable experience of controllers, but to empower them with a tool that handles the complexity of the network—making operations safer, more efficient, and easier to manage.

---

🎥 Project Demonstration

---

🤖 Our Solution: A Practical, Agent-Based AI

Unlike typical ML or RL models, our system is:

• ✅ No Training, No Guesswork → Works from day one.
• ⚡ Ultra-Low Latency → Real-time decision making.
• 🧠 Rules + Logic Driven → Built on Indian Railways regulations.

👥 AI Agents

• Main AI Agent (Chief Controller) → Delegates and supervises.
• Regular Flow Agent → Manages day-to-day operations.
• Train Rerouting Agent → Strategically reroutes trains to minimize delays.
• Accident Handling Agent → First responder in emergencies.
• Priority Management System → Weighted delays (e.g., 1 hr Rajdhani delay = 5 hr effective delay).

---

⚡ Handling Railway Complexity

• 🚨 Accidents & Emergencies → Instantly blocks affected tracks & reroutes trains.
• 🔀 Intelligent Rerouting → Maximizes available track usage while minimizing delays.
• 🎯 Priority Management → Ensures high-priority trains always get preference.

---

📊 Controller Dashboard

• 🗺️ Live Simulation of train movements.
• 🚦 Emergency Highlighting → Blocked tracks in red.
• 🎛️ Manual Handover → Controller always retains full command.

🎬 The dashboard demonstration in the video showcases:

• Real-time movement of trains with color-coded routes.
• Interactive delay injection and agent response.
• Multi-agent workflow (Main, Flow, Rerouting, Accident Handling).
• Performance KPIs and efficiency metrics.

---

📷 Screenshots

---

🏗️ System Architecture

sequenceDiagram
    participant Data as Live Data Input
    participant Main as Main AI Agent
    participant Flow as Flow Agent
    participant Reroute as Rerouting Agent
    participant Accident as Accident Agent
    participant Controller as Human Controller

    Data ->> Main: Receive train positions, delays, incidents
    Main ->> Flow: Manage normal train flow
    Main ->> Accident: Handle emergencies (block tracks)
    Main ->> Reroute: Find optimal new routes
    Accident ->> Reroute: Request rerouting during disruption
    Reroute ->> Main: Return rerouting plan
    Main ->> Controller: Display AI suggestions on dashboard
    Controller -->> Main: Approve / Override decision

Loading