SmartFlow AI Traffic Management System Prototype

SmartFlow: Adaptive AI Traffic Optimization, Computer Vision Vehicle Tracking (YOLO11 + ByteTrack), Priority Emergency Clearance, and SUMO Controller Integration.

---

👥 Team Work Division & Architecture

                 Traffic Video
                      │
                      ▼
      ┌─────────────────────────────────┐
      │ Member 2                        │
      │ Computer Vision Engineer        │
      │ YOLO11 + ByteTrack              │
      └─────────────────────────────────┘
                      │
                      ▼
     Vehicle Count • Queue Length • Speed
                      │
                      ▼
      ┌─────────────────────────────────┐
      │ Member 1                        │
      │ AI Lead & System Integration    │
      └─────────────────────────────────┘
                      │
                      ▼
          Traffic State Information
                      │
                      ▼
      ┌─────────────────────────────────┐
      │ Member 3                        │
      │ Traffic Control Engineer (SUMO) │
      └─────────────────────────────────┘
                      │
                      ▼
      Waiting Time • Delay • Throughput
                      │
                      ▼
      ┌─────────────────────────────────┐
      │ Member 4                        │
      │ Dashboard & Documentation       │
      └─────────────────────────────────┘
                      │
                      ▼
       Final Software Prototype & Demo

---

🚀 Focus: Member 2 – Computer Vision Engineer

📋 Member 2 Responsibilities

1. Traffic Video Collection & Processing: Support MP4, AVI, and MOV traffic video sources.
2. YOLO11 Object Detection: Multi-class vehicle detection (Car, Truck, Bus, Emergency Ambulance).
3. ByteTrack Multi-Object Tracking: Unique persistent ID assignment across video frames.
4. Vehicle Density & Count Extraction: Frame-by-frame vehicle counts per corridor (North, South, East, West).
5. Queue Length Estimation: Meter-based queue occupancy calculation.
6. Speed Estimation: Centroid displacement velocity tracking ($\text{km/h}$).
7. Perception Telemetry Export: Export structured CSV files for Member 1 (AI Lead) and Member 3 (Traffic Control Engineer).

---

🛠 Project Directory Structure

Trafficflow/
│
├── datasets/                            # Member 2: Labeled traffic dataset images/labels
├── videos/                              # Member 2: Input traffic videos (MP4, AVI, MOV)
├── notebooks/                           # Member 2: Jupyter / Colab notebooks
│   └── Member2_YOLO11_ByteTrack_Pipeline.ipynb
├── models/                              # Member 2: Trained YOLO11 weights (yolo11n.pt, yolo11s.pt)
├── outputs/                             # Member 2: Tracked videos & perception CSVs
│   ├── tracked_traffic_video.mp4
│   └── traffic_perception_metrics.csv
│
├── detection/                           # Member 2: Vision & Tracking Engine
│   └── yolo11_bytetrack_pipeline.py
├── tracking/                            # Member 2: ByteTrack configuration
├── feature_engine/                      # Member 1: Traffic State Generator
├── controller/                          # Member 3: Adaptive Max-Pressure Logic
├── sumo/                                # Member 3: SUMO Simulation network
├── src/                                 # Member 4: React Web Dashboard & Canvas
├── requirements.txt                     # Dependencies
└── README.md                            # Documentation

---

⚡ Member 2 Execution Guide

Option A: Run Standalone Python Script (Windows 11 / Linux)

1. Install Dependencies:

   pip install -r requirements.txt

2. Execute YOLO11 + ByteTrack Pipeline:

   python detection/yolo11_bytetrack_pipeline.py --input videos/traffic.mp4 --output outputs/tracked_traffic_video.mp4 --csv outputs/traffic_perception_metrics.csv

---

Option B: Run in Google Colab (GPU Accelerated)

Open notebooks/Member2_YOLO11_ByteTrack_Pipeline.ipynb in Google Colab to run GPU-accelerated YOLO11 inference and download traffic_perception_metrics.csv.

---

📊 Member 2 Output Contract (Structured CSV for Member 1 & Member 3)

The vision pipeline outputs outputs/traffic_perception_metrics.csv containing:

frame_id	timestamp_sec	corridor_lane	vehicle_count	queue_length_m	average_speed_kmh	emergency_detected
30	1.0	NORTH 4-LANE	12	132m	28.5 km/h	0
30	1.0	SOUTH 4-LANE	14	154m	36.2 km/h	1 (AMBULANCE)
30	1.0	EAST 4-LANE	10	110m	30.0 km/h	0
30	1.0	WEST 4-LANE	6	66m	44.1 km/h	0

---

💻 Running the Web Dashboard (Member 4 Interface)

1. Install Node Dependencies:

   npm install

2. Start Local Server:

   npm run dev

   Open http://localhost:5173 in your browser.