🛰️ GridWatch: AI-Powered Drone Fleet Optimization

🚀 Inspiration

The NextEra Energy Drone Optimization Challenge posed a fascinating real-world problem at the intersection of operations research, geospatial analytics, and renewable energy infrastructure.

What inspired us most was the scale and impact of the challenge:

Tens of thousands of miles of electrical grid lines requiring inspection
Autonomous drone fleets operating at massive scale
Critical infrastructure powering millions of homes

This wasn’t just about finding any solution — it was about designing an efficient, scalable, and reliable system that could reduce operational costs while ensuring complete and safe coverage of NextEra Energy’s assets.

📘 What We Learned

🧮 Operations Research Fundamentals

We explored the foundations of Vehicle Routing Problems (VRP) and Capacitated Vehicle Routing Problems (CVRP). Through this, we learned to:

Model complex real-world constraints mathematically
Use Google OR-Tools to find near-optimal fleet assignments
Handle battery capacity and distance limitations per drone mission

🌍 Geospatial Data Processing

We gained hands-on experience working with real-world geographic datasets, including:

Processing WKT polygons for flight zone boundaries
Performing coordinate transformations (longitude ↔ latitude)
Reconstructing optimal paths using Dijkstra predecessor matrices
Writing defensive code to manage out-of-bounds indices gracefully

💻 Full-Stack System Design

We built a complete platform integrating optimization with visualization:

Backend: Python (modular, API-based architecture)
Frontend: React with real-time data visualization
Integration: RESTful communication between frontend and backend
Validation: Interactive Plotly maps to verify coverage and constraints

🧠 Mathematical Optimization

At the core, our objective was to minimize total flight distance while satisfying operational constraints:

Subject to:

Coverage Constraint: Every photo waypoint must be visited
Battery Constraint:
37,725 ft per drone
Airspace Constraint: All paths must remain within designated flight zones

🧩 Frontend–Backend Integration

We created a React dashboard that enables users to:

Configure and monitor drone fleet missions
Send optimization requests to the backend
Track optimization progress through animated status bars
Visualize missions on interactive maps with real-time updates

🗺️ Visualization & Validation

Our system provides clear, data-driven insights through advanced visualization tools:

Required Waypoints (orange dots): Mandatory coverage points
Optimized Paths (colored lines): Drone routes minimizing travel cost
Uncovered Zones (red X’s): Highlighting coverage gaps
Flight Zone Boundaries (red polygon): Ensuring compliance

⚙️ Challenges Faced

1. Constraint Complexity

Balancing coverage requirements with limited battery life and restricted airspace was a major challenge.
Solution: We used a predecessor matrix approach to reconstruct safe, valid paths within polygon boundaries.

2. Data Quality Issues

Some indices in asset_indexes.npy exceeded valid bounds.
Solution: We implemented defensive filtering based on recommendations in the challenge guide.

3. Visualization Design

We needed intuitive visuals to distinguish missions and expose gaps.
Solution: Enhanced our Plotly-based maps with dynamic coloring, coverage checks, and hover-based interactivity.

🏁 Outcome

GridWatch successfully demonstrates how AI-driven optimization and geospatial analytics can transform drone-based infrastructure inspection. By merging mathematical rigor, software engineering, and data visualization, we created a system ready to scale with real-world renewable energy operations.