How to implement Machine Learning using nRF51822 and Accelerometer ADXL345

[mantap]

Hello, somebody help please..
I want to implement Machine Learning using tensorflow, KNN or whatever is good for my module, that consist of Bluetooth Low Energy nRF51822 type and Accelerometer I2C ADXL345. and the methods that I was using before are How to make Accelerometer data that were 3 axis (X, Y, Z). after I can show the raw data from my accelerometer, so I want to make thresholding value for step detection using euclidean norm such as sqrt(x^2+y^2+z^2). after I can measure step value from subject. I want to make distance value, which is from 2x integral from accelearation data to distance that using trapezoidal method that can measure like this one (delay)*(th1+th2)/2. 

So after several works, I want to add some method that can classify the every human activity recognition, so I think using machine learning it's the best way to collect pattern data, like calibration function. I want to use tensorflow. but it can't to implemented because my module have small ram just arround 16kb maybe. so How your suggestion and code guys.. please help me, I'm using mbed compiler to make easy way..

please help..

[Janek Mann]

Machine learning for resource-constrained environments is a rather complicated topic. You can't just expect to take an algorithm developed in Tensorflow and port it across to a Cortex-M0. 

You could implement a model of the resource constraints on the NRF51822 and attempt to train a model within those constraints, but I am not aware of any out-of-the-box solutions for this. 

For human activity recognition there is a decent body of existing research on hand-designed implementations that can run in a microcontroller.

Alternatively, a popular approach is to implement low-power feature extraction on the MCU and use a model in say a phone app to perform the actual recognition. But even if you want to implement the recognition on the MCU the separation of feature extraction and recognition is a useful approach in a resource-constrained environment.

Hope that gives you some useful terms to Google, and don't forget about scholar.google.com too!

[mantap]

thanks for answering, but If you can give me example code to implement this machine learning to my module. it's very usefull..

because I still confuse to separate from main program that recognize some pattern calibrate function program.

please help...