Data Collection System and Stream Learning Model for ADL Classification

Overview

This repository contains the complete implementation of the thesis project:

Data Collection System for Training a Stream Learning Model for Activities of Daily Living (ADL) Classification

The system was designed to support real-time classification of human activities using inertial data collected from smartphones (accelerometer and gyroscope).

The project integrates:

• Offline supervised training
• Online incremental (stream) learning
• Selective label request policy
• Stability and reaction analysis over time

The architecture follows a client–server model with persistent storage and incremental adaptation.

---

Online Server – Real-Time ADL Streaming Backend

Description

This repository contains the Python WebSocket backend deployed on Render as part of the thesis system.

The server is responsible for:

• Receiving real-time windowed sensor data from the Flutter mobile client
• Validating incoming JSON payloads
• Performing optional real-time inference
• Supporting the online Stream Learning workflow

Live Deployment (Render):
https://online-sensor-backend.onrender.com

---

System Role in the Architecture

Client (Flutter App)
→ Streams feature windows via WebSocket
→ Online Server (Render)
→ Prediction + policy logic
→ Logging / evaluation / incremental update

This server represents the online experimentation and deployment layer of the thesis architecture.

---

Key Features

• WebSocket-based real-time communication
• Structured JSON window payloads
• Schema validation (optional)
• Real-time inference (SVM + probability calibrator)
• Selective labeling policy engine
• Incremental evaluation compatibility (River – HT / ARF)
• Cloud-ready deployment (Render)

---

Example Payload (Client → Server)

{
  "device_id": "phone_01",
  "position": "pocket",
  "sampling_hz": 20,
  "window_id": "abc123",
  "timestamp": "2025-08-20T15:32:10Z",
  "features": {
    "acc_mean_x": 0.12,
    "acc_std_x": 0.98,
    "gyro_mean_z": -0.03
  },
  "metadata": {
    "session_id": "s01",
    "user_id": "u01"
  }
}

---

Example Payload (Server → Client)

{
  "status": "ok",
  "window_id": "abc123",
  "prediction": "walking",
  "confidence": 0.82,
  "request_label": false
}

---

Technologies Used

• Python
• Async WebSocket server
• Scikit-learn (offline baseline model)
• River (incremental models – optional integration)
• JSON schema validation
• Render (cloud deployment)

---

Academic Context

This work contributes to research in:

• Real-time Human Activity Recognition (HAR)
• Activities of Daily Living (ADL) classification
• Stream Learning methodologies
• Incremental model evaluation
• Stability and adaptation analysis in evolving systems

---

Author

Paula M.
Electronic and Telecommunications Engineering
Universidad del Cauca

---

License

This repository is intended for academic and research purposes.