Eagle Eye – AI-Powered Port Traffic Optimization

Eagle Eye is an AI assistant for Baltic ports that predicts ship arrivals, flags congestion risks, and helps operations teams schedule berths, pilots, and tugs before delays happen.

Originally built during the Portathon in Klaipėda 🇱🇹, Eagle Eye focuses on reducing idle time at anchor, cutting CO₂ emissions, and giving ports a clear, predictive view of incoming traffic.

---

🌍 Problem

Large vessels waiting at anchor:

• Burn fuel and emit CO₂ while idling
• Cost tens of thousands of dollars per day in delays
• Create uncertainty for terminals, pilots, and hinterland logistics

Unpredictable arrivals → bottlenecks, last-minute firefighting, and inefficient resource use.

---

🚀 Solution

Eagle Eye turns AIS streams and schedule data into actionable forecasts and a chat-style assistant for port operators:

• Predicts ETA and congestion risk for inbound vessels
• Highlights likely bottlenecks at berths, channels, or time windows
• Suggests proactive actions (e.g., rescheduling, prioritization, tug/pilot planning)
• Provides a conversational interface to query “what if?” scenarios

Example questions Eagle Eye can answer

• “Which vessels are likely to arrive in the next 6 hours and cause congestion?”
• “Show me predicted waiting time per berth tomorrow.”
• “If we delay Vessel X by 2 hours, what happens to overall waiting time?”

---

✨ Key Features

• AIS-driven forecasting
  • Ingests AIS data and schedules to predict ETA and traffic peaks.
• AI assistant for port ops
  • Chat-style UI over LLM + RAG to explain forecasts, vessels, and scenarios.
• Congestion risk alerts
  • Flags time windows and areas with likely overload, enabling early action.
• Ops-aware recommendations
  • Helps align berths, pilots, and tugs with forecasted arrivals.
• Deployable as SaaS
  • Designed as a lightweight, containerized service that can be deployed with minimal IT effort.

---

🧱 High-Level Architecture

Data & Models

1. AIS Ingestion Pipeline (Python)

   • Fetches and preprocesses AIS streams and schedule data
   • Cleans, filters, and aggregates vessel tracks and port calls

2. Time-Series Forecasting Engine

   • Predicts arrivals, waiting times, and traffic density over future time windows

3. Knowledge Layer (LLM + RAG)

   • Indexes manuals, port rules, operational guidelines, and historical notes using LangChain + FAISS
   • LLM uses this indexed context to provide grounded, port-aware answers

Services & Interfaces

4. FastAPI Backend

   • REST endpoints for forecasts, vessel states, and chat queries
   • Orchestrates forecasting engine and RAG pipeline

5. React Frontend

   • Web dashboard for:
     • Traffic timeline and congestion hotspots
     • Vessel lists and ETA predictions
     • Chat-style assistant for ops questions

6. Dockerized Deployment

   • Backend, frontend, and supporting services containerized for easy deployment
   • Suitable for on-prem or cloud-hosted setups

---

🛠 Tech Stack

• Backend: Python, FastAPI
• AI & Data: AIS data pipelines, time-series forecasting, LLM + RAG (LangChain + FAISS)
• Frontend: React
• Infrastructure: Docker
• Domain: Port operations, maritime logistics, sustainability

---

📊 Expected Impact (Hackathon Prototype)

Based on Portathon assumptions and scenario modeling:

• ~30% reduction in average waiting time per vessel
• 12,000–15,000 tonnes of CO₂ emissions avoided annually
• $15–20M annual cost savings for a large port
• Return on investment estimated within months for full-scale deployment oai_citation:0‡Eagle_Eye.pptx

Note: These are indicative figures from the hackathon concept phase, not audited production metrics.

---

📂 Repository Contents

Because the original hackathon codebase is not publicly shareable, this repository is intended as a concept and documentation repo, typically containing:

• README.md – Project overview and architecture (this file)
• docs/ – Design notes, mockups, and process documentation
• slides/ – Pitch deck / Portathon presentation (including “How Eagle Eye Works” slide)
• roadmap.md – Ideas for future iterations and pilots

---

🔮 Future Directions

Potential next steps for Eagle Eye:

• Integrate real-time APIs from multiple Baltic ports
• Improve models with weather, tidal, and terminal productivity data
• Add “what-if” simulation tools for planners
• Integrate with existing Port Community Systems (PCS) and terminal software
• Extend sustainability analytics (CO₂, fuel, and cost dashboards)

---

🤝 Acknowledgements

Eagle Eye was created during the Portathon in Klaipėda with guidance from organizers, mentors, and partner ports. Special thanks to Lawrence Henesey, Ph.D. for mentorship, strategic feedback, and support throughout the sprint.

---