A complete IoT project demonstrating alerting and monitoring of an automotive engine in real-time using an ESP32 microcontroller. This system performs on-device inference with a quantized TensorFlow Lite (TFLite) model and streams the results to AWS IoT Core for cloud monitoring.
This project combines embedded systems, machine learning, and cloud computing to detect engine health anomalies. It integrates:
- On-Device inference using a quantized TFLite Model.
- AWS IoT Core MQTT messaging for remote cloud communication.
- A complete ML Pipeline from data generation and model training to deployment on an embedded microcontroller.
- ESP32 development board
- Arduino IDE
- AWS IoT Core Account
- Python 3.x environment for machine learning tasks
You will need to install the following libraries in Arduino IDE:
- Adafruit Unified Sensor 1.1.15
- ArduinoJson 7.3.0
- DHT sensor library 1.4.6
- EloquentTensorFlow32 1.0.4
- EloquentTinyML 3.0.1
- ESPPubSubClientWrapper 1.0.2
- PubSubClient 2.8
- TensorFlowLite_ESP32 0.8.0
- tflm_esp32 2.0.0