Tilt Detection System – Embedded C++ + Node.js Server + Web Interface

A complete system for tilt detection using the MPU6050 sensor with a Raspberry Pi 5.
This project includes:

• C++ embedded code for sensor data acquisition
• Node.js server with REST API
• A static HTML+JS frontend served via Express whose content will then be dinamically updated through WebSocket and JavaScript

---

Repository Structure

main branch: current working version
reactjs-version branch: previous version with React interface

---

Development Environment

• Host machine: Ubuntu 24.04 (amd64)
• Target machine: Raspberry Pi 5 (Debian 12 – arm64)
• IDE: Visual Studio Code

---

Toolchain

• Compilers: aarch64-linux-gnu-gcc/g++-12
• Libraries: i2c, curl, pthread
• Build system: CMake

---

Embedded Code (Example.cpp)

• Language: C++
• Description:
  • Initializes MPU6050 (3 axys accelerometer + 3 axys gyroscope)
  • Computes offsets
  • Reads raw sensor data and calculates tilt angles
  • Packs data into a JSON object
  • Sends it to the server using a POST HTTP request via curl

---

Backend + Frontend

Stack

• Languages: JavaScript (Node.js + Express), HTML/CSS
• Runtime: Node.js
• Security: TLS encryption over HTTPS
• Real-time updates: socket.io (bidirectional, real-time and low latency communication)

Communication Flow

C++ → Node.js (POST)

• The embedded client sends a JSON payload to /data

Node.js → Browser (WebSocket)

• The server forwards real-time data to connected clients via socket.io. Data contained in the chart and table will update every time a new JSON payload is received.

Browser → Node.js (GET)

• The browser sends a GET request to / and receives the HTML page index.html
• The page includes:
  • A table displaying all received tilt data
  • A live-updating chart
  • Filter by ID working on both table and chart
  • Visual alerts when thresholds (2.5 °) on roll pitch and yaw values are exceeded

---

Web page preview

A preview of what the user would see on his browser when visiting the web page: