Arduino Nano 33 BLE Sense for a TinyML is a most popular board to demonstrate on-device learning and inference as it has lot of sensors on the board. One of the case which can be executed on this board is a "Linear Trend predictor" where using the onboard temperature sensor, we will be predicting the temperature for the next 5 days.
We are trying to use TinyML independently and we will see how it goes. The purpose of this experiment is that we will build a small model that learns from the past few temperature readings and predicts the next 5 days temperatures.
- For this we need temperature data of the past for the model to get trained, if we do not have the past data, we can collect the current data for reference
- The other steps include
- defining the model (on your laptop or desktop)
- training the model (on your laptop or desktop)
- converting the model so that it can imported to your board (on your laptop or desktop)
- Create Arduino sketch for inference
]]>
0 Comments