This post covers how ARNIE can be controlled by using an Xbox Kinect peripheral.

Xbox Kinect

This post details the core component of ARNIE’s brain, the Spine software, and how the networking system for the robot works.

This post provides an overview of the entire ARNIE architecture, with the intention of providing a baseline understanding for how everything works together before I dive into the deep technicals of it all.

An introduction to this series

Since October I have been working in a team of 4 students in the Cybernetics department at the University of Reading developing a robot called ARNIE as part of our degree. The MEng Challenge is a 30-credit (25% of our year’s total) module where we work in a team to design and build a system to achieve a certain goal.

I have been sharing bits of information about ARNIE as we have been developing it, but I thought it was time to get into the gory details and blog in detail about what we’ve been up to. This is the first in a series of blog posts I have produced that explain how the robot works overall, then dives into details about the particular aspects I have been working on.

I have written and talked in the past on the subject of hacking Kinect, and how awesome it would be to hook it up to our robot project at Uni. Well I’m pleased to say that last week we achieved robot arm control with the Kinect sensor, using it to measure arm position and relay it to the arm, which then moved to mirror the action.

Check out the videos below (instructional and cheesy each in their own way) for demos of the Kinect and arm doing its thing.

Credits go to Stefan, Jean-Jacques and Ioannis (my fellow team members), and our supervisor Prof. Harwin for letting us play around with this stuff. If you want the code, it’s all free on Github here and here.