ESP_ML

ESP_ML is a lightweight client–server framework designed to enable real-time sensor data collection from ESP32 microcontrollers and seamless communication with cloud-hosted machine learning models via REST APIs.

It bridges the gap between embedded IoT systems and cloud-based ML inference, providing a scalable foundation for intelligent IoT applications.

Project Website: https://esp-ml-backend.onrender.com/

---

Overview

ESP_ML consists of two core components:

1. ESP32 Client Library (C/C++)

   • Collects real-time sensor data.
   • Communicates with the backend ML inference API using HTTP/REST.

2. ML Backend Server (Python + Flask)

   • Hosts trained ML models.
   • Manages incoming sensor data and performs inference.
   • Sends results back to connected ESP32 clients.

---

Project Structure

ESP_ML/
├── esp_ml_client/           # ESP32 client source code (C/C++)
│   ├── esp_ml_client.cpp
│   ├── esp_ml_client.h
│   └── library.properties
│
├── ml_backend/              # Python backend service
│   ├── api.py               # Flask API endpoints
│   ├── db.py                # Local/SQLite database handler
│   ├── model_manager.py     # Model loading & inference logic
│   ├── utils.py             # Helper utilities
│   ├── requirements.txt     # Python dependencies
│   ├── templates/           # Web UI templates
│   ├── static/              # CSS, images, and client zip
│   └── keys.db              # Local key store (optional)
│
└── README.md

---

Installation & Setup

Backend (Flask ML Server)

Prerequisites

• Python 3.10+
• pip package manager

Setup

cd ESP_ML/ml_backend
python -m venv venv
source venv/bin/activate   # On Windows: venv\Scripts\activate
pip install -r requirements.txt

Run the Server

python api.py

Default address: http://127.0.0.1:5000 You can host the backend on a cloud VM, Raspberry Pi, or any REST-accessible environment.

---

ESP32 Client

Prerequisites

• Arduino IDE or PlatformIO
• ESP32 board package installed
• Wi-Fi network credentials
• URL of the ML backend server

Integration Steps

1. Include the ESP_ML client library in your ESP32 project.
2. Initialize the client with your Wi-Fi credentials and backend URL.
3. Send sensor data and receive inference results using the provided API.

Example Sketch

#include "esp_ml_client.h"

ESPMLClient client("your_ssid", "your_password", "http://<server_ip>:5000");

void setup() {
  Serial.begin(115200);
  client.connectWiFi();
}

void loop() {
  float sensorValue = analogRead(34);
  String result = client.sendData(sensorValue);
  Serial.println("Prediction: " + result);
  delay(2000);
}

---

Features

✅ Real-time sensor data acquisition
✅ Lightweight REST communication for embedded systems
✅ Cloud-hosted ML inference via Flask
✅ Modular architecture — replaceable backend or model
✅ Easy ESP32 integration
✅ Optional visualization dashboard (HTML templates)

---

Backend API Endpoints

Endpoint	Method	Description
/predict	POST	Receives JSON sensor data and returns ML inference
/register	POST	Registers a new device/client
/status	GET	Health check endpoint
/upload_model	POST	Uploads or replaces an ML model (admin use)

Example Request

POST /predict
{
  "device_id": "esp32_01",
  "data": [0.21, 0.34, 0.56, 0.12]
}

Example Response

{
  "prediction": "normal",
  "confidence": 0.92
}

---

Dependencies

Backend (Python)

Listed in ml_backend/requirements.txt:

Flask
numpy
pandas
scikit-learn
requests

(Dependencies may vary depending on your model setup.)

ESP32 Client (C/C++)

• Arduino Core for ESP32
• WiFiClient / HTTPClient (built-in)
• ArduinoJson

---

Configuration

Component	Parameter	Description
Client	SSID, PASSWORD	Wi-Fi credentials
Client	API_URL	URL of backend ML API
Backend	MODEL_PATH	Path to trained ML model
Backend	PORT	Flask server port

---

Example Workflow

1. Deploy Flask backend (local/cloud).
2. Connect ESP32 to Wi-Fi and set backend URL.
3. ESP32 sends real-time sensor readings to the API.
4. Backend performs ML inference (e.g., anomaly detection).
5. ESP32 receives and displays prediction results.

---

Troubleshooting

Issue	Cause	Solution
ESP32 not connecting to Wi-Fi	Incorrect SSID/PASSWORD	Verify credentials
No response from server	Wrong API URL or port	Check Flask server IP and port
Connection refused	Server inactive/firewall issue	Ensure Flask is running, port 5000 open
Unexpected inference output	Model or data mismatch	Verify JSON format and input schema

---

Contributor

M Sreenath Karthick — Developer

Contributions are welcome! Feel free to open issues or submit pull requests to improve the project.