I need an opinion on using Edge impulse and Thunderboard sense 2 in my project.
So my project is to put the thunderboard in a ball, powered by a cell battery. Before that i have to acquire data through board sensors, practice machine learning to label 3 movements with the thunderboard in the ball. And then implement this model in the board to use it to see in my phone the movement that has been made with the ball (knowing that the thunderboard is flashed and is only powered with a cell battery) through BLE.
Is working with edge impulse and thunderboard a good idea ?
If you’re trying to classify movements with the ball using the Thunderboard, then Edge Impulse is a good fit, as it will help you train and deploy your machine learning model.
I don’t know how long that battery will last, though. Remember that performing inference to classify motion (or anything) requires your CPU to be running constantly. You might be able to create some kind of interrupt that wakes the CPU on motion, have it capture the motion data for some period, classify it, and then send it to your phone. That might help save on battery life.
Also, while Edge Impulse will help you create code to perform machine learning on your Thunderboard, we don’t help with BLE communication. You will likely need to find another library to help you with that.
Hope that helps!