CareRoute 🚑

Smart Ambulance Decision-Support System

Recommending the most suitable hospital during emergencies — not just the nearest one.

---

What is CareRoute?

Most ambulances go to the nearest hospital. CareRoute goes to the right one.

It evaluates every nearby hospital in real time based on patient vitals, ICU availability, specialist match, live traffic, and predicted bed availability — then recommends the best option with a full AI-generated explanation of why.

---

Features

• 🧠 AI Reasoning Engine — Explains why a hospital was selected and why others were rejected
• 🔄 Autonomous Rerouting — Detects mid-route capacity changes and switches destination automatically
• 🎙️ Voice-Captured Vitals — Paramedic speaks vitals hands-free, system transcribes in real time
• 📍 Distance & travel-time calculation — Live traffic-aware routing
• 📊 Predictive bed availability — Forecasts capacity 15 minutes ahead
• 📡 Ambulance-to-hospital communication — Pre-alert sending before arrival

---

Tech Stack

Layer	Technology
Frontend	React.js
Backend	Node.js, Express
AI Reasoning	OpenAI API
Voice Transcription	OpenAI Whisper (gpt-4o-mini-transcribe)
Database	JSON (simulation)
Communication	EventEmitter
Containerization	Docker, Docker Compose
Orchestration	Kubernetes (Minikube)

---

Project Structure

smart-ambulance/ │ ├── frontend/ # React dashboard UI │ ├── components/ │ │ ├── Vitals.js # Patient vitals display │ │ ├── HospitalList.js # Ranked hospital recommendations │ │ ├── MapView.js # Live route map │ │ └── Buttons.js # Emergency controls │ ├── App.js │ ├── api.js # API integration layer │ └── Dockerfile │ ├── backend/ # Node.js server │ ├── server.js │ ├── routes.js │ ├── simulation.js # Hospital capacity simulation │ └── Dockerfile │ ├── logic/ │ └── recommendation.js # Scoring & ranking engine │ ├── database/ │ └── hospitals.json # Hospital dataset │ ├── k8s/ # Kubernetes manifests │ ├── backend-deployment.yaml │ ├── backend-service.yaml │ ├── frontend-deployment.yaml │ └── frontend-service.yaml │ ├── docker-compose.yml └── README.md

---

Setup & Installation

Option 1: Run locally (no containers)

1. Install dependencies

npm install

2. Start the backend (Terminal 1)

cd backend
node server.js

Backend runs at: http://localhost:5000

3. Start the frontend (Terminal 2)

cd frontend
npm install
npm start

Frontend runs at: http://localhost:8000

Option 2: Run with Docker Compose

docker-compose up --build

This builds and runs both frontend and backend containers together, with the same port mapping as above.

---

Deployment (Docker + Kubernetes)

CareRoute's backend and frontend are containerized with Docker and can be deployed on Kubernetes for scalable, self-healing orchestration instead of running as single processes.

Why Kubernetes here: running multiple replicas of each service means one crashing pod doesn't take down the app — Kubernetes automatically restarts it. It also makes the deployment declarative (the desired state lives in YAML, not in manual terminal commands) and mirrors how this stack would run in a real production environment on a managed cluster (EKS/GKE/AKS) instead of just a laptop.

Prerequisites

• Docker Desktop
• Minikube
• kubectl

Steps

1. Start Minikube

minikube start --driver=docker

2. Build images inside Minikube's Docker environment

minikube docker-env | Invoke-Expression   # PowerShell
# or: eval $(minikube docker-env)          # bash/zsh

docker build -t careroute-backend:latest -f backend/Dockerfile .
docker build -t careroute-frontend:latest ./frontend

3. Apply the Kubernetes manifests

kubectl apply -f k8s/backend-deployment.yaml
kubectl apply -f k8s/backend-service.yaml
kubectl apply -f k8s/frontend-deployment.yaml
kubectl apply -f k8s/frontend-service.yaml

4. Verify pods are running

kubectl get pods

You should see 2 backend and 2 frontend pods in Running state.

5. Access the app

minikube service careroute-frontend-svc --url

Open the printed URL in your browser.

Architecture

Component	Replicas	Port	Service Type
Backend	2	5000	NodePort
Frontend	2	8000	NodePort

Known limitation

The frontend currently points to the backend via a directly configured URL rather than service-discovery/env-based config, since the frontend runs client-side in the browser and can't resolve internal Kubernetes DNS names. In a production setup, this would be solved with an Ingress controller and a proper external DNS name, or by injecting the backend URL at build/runtime via environment variables. This is documented here as a deliberate next step rather than an oversight.

---

Voice Input Setup

Voice input runs in demo mode by default — no API key required for presentations. The browser records a short clip, the backend returns a sample transcript, and parsed values auto-fill the vitals dashboard.

For real speech-to-text, create a .env file in the root directory: OPENAI_API_KEY=your_api_key_here OPENAI_TRANSCRIPTION_MODEL=gpt-4o-mini-transcribe

Optional environment variables:

# Force demo mode even when API key exists
VOICE_TRANSCRIPTION_MODE=mock

# Override the demo transcript
MOCK_VOICE_TRANSCRIPT="Patient is unconscious, heart rate 130, oxygen level 82, blood pressure 90 over 60, temperature 101, possible cardiac emergency."

Note: Microphone access requires localhost or an HTTPS deployment in modern browsers.

How to use: Click Start Voice Capture → speak patient vitals → click Stop and Process (or wait 10 seconds). Recognized values update the dashboard automatically.

---

How the Recommendation Engine Works

Each hospital is scored based on:

• ICU and ventilator availability
• Specialist match for the patient's condition
• Distance and travel time
• Live traffic conditions
• Predicted bed availability (next 15 minutes)
• Current emergency load

The highest-scoring hospital is recommended with a full explanation. Rejected hospitals show specific reasons — full capacity, missing specialty, high traffic, or lower suitability score.

---

API Reference

Recommend a hospital: POST /api/hospitals/recommend

{
  "severity": "CRITICAL",
  "requiredSpecialty": "cardiology",
  "location": {
    "lat": 28.61,
    "lng": 77.20
  }
}

---

Adding Your Own Hospital Data

Edit database/hospitals.json. Each record should follow this format:

{
  "name": "City Heart Hospital",
  "location": { "lat": 28.61, "lng": 77.20 },
  "icu_total": 20,
  "icu_available": 5,
  "ventilators_total": 10,
  "ventilators_available": 3,
  "specialists": {
    "cardiologist": 2,
    "neurologist": 1,
    "trauma": 3
  },
  "status": "available",
  "bedTrendPer15Min": 0
}

Restart the backend after updating the file.

---

Future Scope

• Google Maps / real routing API integration
• Live hospital API connections
• Voice-to-vitals automatic dashboard sync
• Wearable & IoT vitals integration
• Multi-city deployment
• Ayushman Bharat Digital Mission integration
• Ingress-based routing to remove hardcoded backend URL
• CI/CD pipeline for automated build and deploy to Kubernetes

---

Team

AlgoAllies — Built for hackathon, designed for real-world impact.

---

For academic and hackathon use only.