This project is the continuation of previous work on an environmental monitor I designed. It adds AI functionality to the data processing and results reporting part of the boards firmware and software.

The board itself is driven by an STM32 mcu and contains a humidity, temperature and light sensor. The board makes use of the on chip ADC to provide a digital output from each analog sensor.

The software workflow includes:
1. Gather the digital sensor data from the board in csv format. Use the previously attained data to train the AI model.
2. This data is then used to train and create a model on a PC or cloud provider platform.
3. The model is optimised and quantized in order to reduce its size and complexity.
4. After these operations have been carried out and refined the model can be uploaded to the board to provide an enhanced ouput.

Compared to a purely digital data stream the AI output will be user friendly and selective in its nature, providing useful information to an end user. It could be used in a consumer electronics setting.  It is envisaged to use Python to train, refine and infer the model. The software will use micropython, C or a combination of both for the firmware uploaded to the mcu.