Ghost Route

Navigation for a World Where the Map Keeps Changing

Most navigation systems answer one question:

“What is the fastest route from A to B?”

But during an emergency, the fastest route right now may become the worst route moments later.

A flooded road, blocked intersection, failed bridge, sudden congestion, or infrastructure disruption can turn an optimal route into a dead end.

Ghost Route is a resilience-aware navigation system that finds routes designed to remain viable as conditions change.

---

Inspiration

We were inspired by a simple problem: in an emergency, reaching the destination matters more than choosing the shortest path.

Traditional routing generally optimizes for factors such as distance or travel time. However, emergencies introduce uncertainty. Roads can become inaccessible, hazards can spread, and failures in one part of a network can create cascading effects elsewhere.

We wanted to explore a different question:

What if navigation optimized not only for speed, but for survivability?

That question became Ghost Route.

---

What It Does

Ghost Route models a transportation network as a dynamic graph and evaluates routes using multiple factors, including:

• Travel time
• Hazard exposure
• Congestion
• Road reliability
• Infrastructure vulnerability
• Network redundancy
• Potential cascading failures

Instead of returning only the shortest path, Ghost Route evaluates candidate routes and identifies the route that provides the best balance between speed, safety, and resilience.

The underlying route score can be represented as a weighted multi-objective function:

$$ Score(R) = w_tT + w_hH + w_cC + w_vV - w_rR $$

where the terms represent travel time, hazard exposure, congestion, vulnerability, and route resilience.

The weighting changes according to the scenario.

In normal conditions, travel time can be prioritized.

When a disaster develops, avoiding vulnerable infrastructure and maintaining route survivability becomes significantly more important.

---

Break the City

The centerpiece of Ghost Route is Break the City, an interactive network-disruption simulation.

Instead of treating the city as static, users can progressively introduce failures and observe how the routing network responds.

For example:

NORMAL NETWORK

→ Flood a major road

→ Ghost Route recalculates

→ Disable a critical bridge

→ Traffic shifts

→ Congestion propagates

→ Network resilience decreases

→ A new resilient route emerges

This demonstrates an important property of real transportation networks:

a local failure can produce system-wide consequences.

Ghost Route makes those consequences visible and allows users to see how routing decisions evolve as the network deteriorates.

---

How We Built It

We intentionally separated intelligent interpretation from deterministic decision-making.

The system does not rely on an AI model to randomly decide which road a person should take during an emergency.

Instead, the routing engine uses deterministic graph-based algorithms and explicit scoring rules. This makes the results reproducible, testable, and explainable.

Backend

The backend was built using:

• Python
• FastAPI
• NetworkX
• Graph-based routing
• Hazard modeling
• Scenario simulation
• Route scoring
• Resilience analysis
• REST APIs

The transportation network is represented as a graph:

• Nodes represent intersections and infrastructure points.
• Edges represent road connections.
• Edge attributes capture factors such as travel time, capacity, hazards, and reliability.

When conditions change, the graph is updated and candidate routes are recalculated.

Frontend

The frontend was built using:

• Next.js
• React
• TypeScript
• Interactive vector-based mapping
• Route comparison
• Scenario controls
• Resilience metrics
• Network simulation visualizations

The interface was designed as an emergency operations dashboard, rather than a conventional navigation interface.

---

Challenges

Modeling a Dynamic Network

Real cities contain millions of road connections, vehicles, infrastructure dependencies, and unpredictable events.

For a hackathon prototype, we needed a network that was small enough to simulate interactively while still demonstrating meaningful failure propagation.

We addressed this by creating a structured demonstration network with explicit road, hazard, scenario, and resource data.

Making Routing Explainable

A route recommendation is more useful when users understand why it was selected.

Ghost Route therefore exposes the factors behind its decisions, including travel time, risk, resilience, and critical network segments.

Balancing Speed and Resilience

The fastest route is not necessarily the most reliable route.

A path that saves several minutes but depends on one vulnerable bridge may be significantly worse than a slightly slower route with multiple alternatives.

This led us to treat routing as a multi-objective optimization problem rather than a simple shortest-path calculation.

---

What We Learned

The biggest lesson was that resilience is different from efficiency.

A transportation network can be extremely efficient under normal conditions while being highly fragile during disruption.

We learned to think beyond:

“How quickly can I get there?”

and instead consider:

“How likely is this route to remain usable?”

We also learned that resilience cannot be represented by a single factor. Travel time, hazard exposure, infrastructure vulnerability, redundancy, and cascading effects all interact.

---

Impact

Ghost Route is designed around a broader principle:

Transportation infrastructure should be optimized not only for normal conditions, but for the moment when normal conditions disappear.

The same approach could support:

• Emergency evacuation
• Ambulance routing
• Fire and rescue operations
• Flood response
• Disaster logistics
• Critical supply delivery
• Urban resilience planning
• Infrastructure risk analysis

With access to real-world data, the system could incorporate live traffic, weather, flood information, road closures, infrastructure data, and historical incidents.

---

What's Next

The current prototype demonstrates the core resilience engine through a controlled network and simulated disruptions.

The next step would be connecting Ghost Route to live city data so that the network can continuously adapt to real-world conditions.

Our long-term vision is to build a resilience intelligence layer for transportation networks.

Because when the world changes in seconds, the best route isn't necessarily the one that is fastest now.

It's the one that still works when you get there.

Built With

• artificial-intelligence
• data-visualization
• disaster-response
• emergency-response
• fastapi
• graph-algorithms
• networkx
• next.js
• python
• react
• rest
• typescript
• urban-resilience