A little background...

I am Utkarsh Dwivedi, an undergraduate at IIT Guwahati pursuing Bachelors in Design. This blog is my attempt to build a gesture recognizer that maybe placed on the hand preferably.

I have so far studied HMMs, Moving least squares technique for classifying gestures based on accelerometer data. Moving forward I am setting up an accelerometer + arduino, then a acc + arduino + BT.

The larger part of this exercise is to familiarize myself with arduino, accelerometers, HMMs based machine learning, in order to prepare myself better for my Bachelor Thesis Project on Wearable Computing.

I would request all viewers of this project to help me improve my blogging skills and the technical stuff too.
I would like to thank Mr. Keyur Sorathia and EILab IIT Guwahati for all the support and devices provided.

Short plan - setting up

Short Plan

1. arduino + accelerometer
2. arduino + BT + accelerometer

1. Setting up arduino and accelerometer

Parts used : Arduino Duemilanove and Sparkfun ADXL345

Method
Here is the Fritzig sketch, used the ADXL library code for getting the data

This arrangement allows I2C communication which is said to be slower than SPI. My attempt to run the Sparkfun SPI tutorial was sour, but this worked here is the blog that helped. Ignoring the added voltage divider will also work.

So far my Arduino BT-06 has not been working nicely, on mu Ubuntu 14.04, I am having issues pairing it. Now I will try SPI, whatever works will be made portable. Then focus will be on data collection and handling.

2. Arduino + BT
Having difficulties pairing it to my lappy, or else I will have to make do with a circuit connected to my PC.

Starting off

Done with setting up Openframeworks, Eclipse, Arduino in Linux x64 Trusty Tahr.
oF considerations
- run compilePG.sh, to get the project generator tool, that can be included in eclipse as an external tool.

An initial to-do and a general aim of the project to be added...