This project is a comprehensive demonstration of various capabilities available within the Zephyr RTOS ecosystem. It showcases a rich user interface built with LVGL, real-time sensor data visualization, machine learning inference with TensorFlow Lite, and IoT connectivity via MQTT.

✨ Features

The demo uses the LVGL library to create different tabs providing the following features:

• Boot-up Window: A splash screen shown on device startup.
• LVGL User Interface: A responsive, tab-based UI for navigating different features.
• Real-time Data Graphing: Visualizes Inertial Measurement Unit (IMU) sensor data on a live chart.
• IoT Connectivity:
  • Network: Establishes a network connection (Ethernet).
  • MQTT: Sends and receives data (e.g., from UI sliders) to and from an MQTT broker.
• Machine Learning: Performs gesture recognition using a TensorFlow Lite model that takes IMU data as input.
• Device Configuration: A dedicated tab to display system and network configuration.

🎯 Setup Zephyr Environment

First, ensure you have a working Zephyr development environment. If you don't, please follow the official Zephyr Getting Started Guide.

• Zephyr version: 3.7.1 (latest LTS version)
• Zephyr SDK: v0.16.8

Make sure you are using the correct Zephyr version for this project:

# Navigate to your Zephyr installation
cd zephyrproject

# Checkout the correct branch and update dependencies
git switch v3.7-branch
west update

• hash commit: c291ed46f915cb72aef28479c709356f18f37462 (24 January 2025)
• Tested boards:
  • stm32f746g_disco

🛠️ Build

# From the project's root directory
west build -b stm32f746g_disco

⚡️ Flash

west flash

🔧 Project Structure

display-poc-1_jero/
├── CMakeLists.txt      # Main CMake build script
├── prj.conf            # Kconfig project configuration
├── README.md           # This file
└── src/
    ├── main.c              # Main application entry point and thread setup
    ├── config.h            # Project-wide configuration (MQTT, etc.)
    ├── config_network.c    # Network initialization logic
    ├── config_network.h
    ├── jeroagullo_lvgl/    # LVGL UI source files
    |   ├── jeroagullo_tabs.c   # UI tabs implementation
    |   ├── jeroagullo_charts.c # Charting logic for sensor data
    |   ├── jeroagullo_styles.c # Custom LVGL styles
    |   └── ...
    └── mqtt
        ├── jeroagullo_mqtt.c
        └── jeroagullo_mqtt.h

Here's the README.md section for your MQTT configuration:

🔧 MQTT Configuration

The project uses Zephyr's Kconfig system for flexible MQTT configuration. All MQTT settings can be customized without modifying source code.

📋 Configuration Options

The following MQTT parameters are configurable via Kconfig:

Setting	Kconfig Variable	Default Value	Description
Client ID	CONFIG_MQTT_CLIENTID	"my-client"	Unique identifier for this MQTT client
Broker Address	CONFIG_MQTT_SERVER_ADDR	"10.42.0.1"	IP address or hostname of MQTT broker
Broker Port	CONFIG_MQTT_SERVER_PORT	1883	MQTT broker port (typically 1883)
Topic	CONFIG_MQTT_TOPIC	"telemetry"	Default topic for publishing data
Authentication	CONFIG_MQTT_USE_AUTH	y	Enable/disable username/password auth
Username	CONFIG_MQTT_USER_NAME	"my-auth-id-1@my-tenant"	MQTT authentication username
Password	CONFIG_MQTT_PASSWORD	"my-password"	MQTT authentication password

⚙️ Configuration Methods

Method 1: Edit prj.conf (Recommended for Development)

# MQTT Configuration
CONFIG_MQTT_CLIENTID="stm32-sensor-01"
CONFIG_MQTT_SERVER_ADDR="10.42.0.1"
CONFIG_MQTT_SERVER_PORT=1883
CONFIG_MQTT_TOPIC="sensor-data"
CONFIG_MQTT_USE_AUTH=y
CONFIG_MQTT_USER_NAME="device-user"
CONFIG_MQTT_PASSWORD="secure-password"

Method 2: Interactive Menuconfig

west build -t menuconfig

Navigate to: MQTT Configuration → Modify settings → Save

Method 3: Command Line Override

west build -- -DCONFIG_MQTT_SERVER_ADDR=\"192.168.1.100\" -DCONFIG_MQTT_CLIENTID=\"custom-client\"

🔐 Authentication Handling

• When CONFIG_MQTT_USE_AUTH=n: No credentials sent to broker
• When CONFIG_MQTT_USE_AUTH=y: Username/password sent if provided
• Empty username/password strings are treated as no authentication

🧪 Testing with Mosquitto

For local testing and development, we provide a complete Mosquitto MQTT broker setup.

📖 See detailed Mosquitto setup instructions here: mosquitto/README.md

🚨 Troubleshooting

• Connection refused: Verify Mosquitto is running and firewall allows port 1883
• Authentication failed: Check username/password in Mosquitto configuration
• Network unreachable: Ensure STM32 and PC are on same network subnet

Here's a concise README section about the MPU6050 accelerometer integration:

📊 MPU6050 Accelerometer Integration

The project includes an MPU6050 IMU sensor for real-time motion data acquisition and visualization.

🔌 Hardware Connection

The MPU6050 connects to the STM32F746G-DISCO via I2C1:

STM32F746G-DISCO   ↔   MPU6050
─────────────────────────────────
  I2C1_SCL (PB8)   ↔     SCL
  I2C1_SDA (PB9)   ↔     SDA
      3.3V         ↔     VCC
      GND          ↔     GND

Address Configuration:

• 0x68 (default): AD0 pin connected to GND
• 0x69: AD0 pin connected to VCC

⚙️ Software Configuration

The sensor is defined in the device tree overlay:

&i2c1 {
    imuSensor: mpu6050@68 {
        compatible = "invensense,mpu6050";
        reg = <0x68>;
        status = "okay";
    };
};

🔧 Required Configuration

Add to prj.conf or board config file (like in this project):

CONFIG_I2C=y
CONFIG_SENSOR=y

The accelerometer data is automatically sampled and displayed in the sensor tab of the application, providing real-time motion visualization.

For advanced Mosquitto configuration, security settings, and Docker deployment options, refer to the complete mosquitto/README.md.

📅 TODO list

• Add basic interface in LVGL with styles
• disable/enable the threads when a tab is pressed
• Add mutex to all lvgl functions to synchronize with lv_task_handler() call
• Add Network connection (tested with ethernet)
• Add sliders to corresponding windows
• Add mqtt to send the sliders values
• Add IMU sensor
• Add TensorFlow Lite model
• Add Config tab for configuration

🌟 How to Contribute

Star the repo if it helps you! ⭐

Open an Issue for questions or suggestions. Submit a Pull Request to add new features.