OptiTraffic AI Microservices

OptiTraffic is a microservices-based smart traffic control demo powered by real-time Kafka streaming, deep learning (LSTM), and CityFlow traffic simulation. It features secure JWT-based APIs and a React-based dashboard for visualization and monitoring.

---

🧱 Repository Structure

docker/                  → Docker Compose setup & API Gateway  
microservices/           → Flask-based microservices  
├── login_service/       → Issues JWT tokens for authentication  
├── traffic_signal/      → LSTM-based green light predictions  
├── traffic_monitoring/  → Aggregates and analyzes traffic data  
├── notification/        → Sends traffic alerts via email  
├── simulator/           → WebSocket service for frontend replay  
cityflow_simulation/     → CityFlow runner producing Kafka messages  
ai_model/                → LSTM training and preprocessing scripts  
frontend/                → React dashboard to monitor traffic  

---

🚀 Quick Start

1. Install Docker and Docker Compose.

2. Launch all services from the docker/ directory:

   docker-compose up --build

3. Access the frontend at http://localhost:3000.

4. Use the login form to get a JWT (credentials are in microservices/login_service/login_api.py).

5. The API Gateway listens on port 8080 and proxies requests to internal services.

---

🔐 Microservice Overview

🟢 login_service

• Path: microservices/login_service/login_api.py
• Purpose: Issues JWT tokens on login.
• Endpoint:

  @app.route('/api/login/', methods=['POST'])
  def login():
      ...

---

🚦 traffic_signal

• Path: microservices/traffic_signal/traffic_signal_predictor.py
• Purpose: Predicts green-light road using LSTM or heuristic based on real-time Kafka traffic snapshots.
• Logic:

  if current_mode == 'auto' and lstm_model is not None and len(snapshot_buffer) == SEQUENCE_LENGTH:
      green_light_road = predict_green_light_with_lstm(snapshot_buffer)
  else:
      green_light_road = heuristic_predict_traffic_light(summary)

---

📊 traffic_monitoring

• Path: microservices/traffic_monitoring/traffic_monitoring.py
• Purpose: Builds analytics like average speed, density, and green-light distribution from raw and predicted data.

---

📧 notification

• Path: microservices/notification/notification.py
• Purpose: Subscribes to traffic alerts via Kafka and sends email notifications for accidents or congestion.

---

🔁 simulator

• Path: microservices/simulator/simulator.py
• Purpose: Hosts WebSocket endpoints for replaying simulation data in the frontend dashboard.

---

🧠 AI Model

LSTM model training and preprocessing scripts live in ai_model/:

🧪 Preprocessing & Training

• Preprocessing:
  Converts simulation logs to NumPy arrays

  python preprocess_data.py

• Training:
  Trains a Sequential LSTM model for green-light prediction

  model = Sequential([
      LSTM(32, input_shape=(sequence_length, input_dim), return_sequences=False),
      Dense(16, activation='relu'),
      Dense(num_roads, activation='softmax')
  ])
  model.fit(X_train, y_train_cat, epochs=20, batch_size=32, validation_data=(X_test, y_test_cat))

• The final model (traffic_signal_model.h5) is consumed by traffic_signal.

---

🛡️ Tech Stack

• Microservices: Flask + REST
• Message Queue: Kafka
• Simulation Engine: CityFlow
• ML: LSTM (Keras + TensorFlow)
• Authentication: JWT
• Frontend: React
• DevOps: Docker, Docker Compose
• Security: API Gateway with token validation
• Realtime: WebSocket streaming for simulation