The start of my third year

Well it's been a week since I posted about my Christmas Ale making and I thought I'd post about what's happened since then. The main thing being that I've now officially started my third year.

What do I study?

So, something I've not mentioned so far on my blog is what it actually is I do! I am of course a student as some will have noticed on the front page of this blog:

This is my personal blog, filled with ramblings of a student.

I am an undergraduate student at the University of Reading, working toward an MEng in Artificial Intelligence (A.I.). I am now in my third year and it is expected for me to achieve a 2.1 or first class honours as my final degree classification (of which I very much hope I get the latter).

What is artificial intelligence anyway?

A very broad subject! In order to answer such a question I would have to ask at least one, for example, what do you consider to be intelligence? - This could change what you think artificial intelligence is.

Examples of A.I.

Aritifical intelligence is used in a lot of applications: in quality control, for verification of products on a production line; predicting the weather, by learning patterns in past weather and analysing meteorological data. It's also used in Chat bots, Google's suggestion of searches, feature tracking in video editors and a whole number of other things.

There are of course the more interesting examples of the usage of A.I. most of which have very active research fields in the modern day, these are things like self driving cars or autonomous unmanned aerial vehicles (UAVs).

So come on then, what actually is it?

Looking at all the things considered A.I. it's probably best to say that A.I. is anything you want it to be. Any system that appears to be making choices could be considered A.I. - Facebook when it suggests friends is a form of A.I.

What A.I. is not: A.I. is not simply a mimicry of human intelligence in humanoid form! There are those who say this is what it should be and yes, it is what every A.I. researcher would like to accomplish. However, we're just not there yet.

A.I. that I work with

During my course I've looked at A.I. in many different forms. I've programmed artificial neural networks (ANNs) and this year I shall be implimenting A.I. in terms of control for an autonomous UAV.

Third year project

In the third year on a masters degree course there is a group project. For me and the group I am involved with, our project is called "The flying brain". The aim of the project is to develop control software for a UAV capable of controlling it to navigate through a hallway and a fixed size opening (e.g. a doorway). The UAV will have a single fixed point camera.

Which drone to use?

For simplicity and because the department already has them the group shall be using the Parrot AR.Drone. However we have contacted some other companies about their drone and this may well change as time goes on.

How are we going to do it?

That is the big question! We are going to impliment Visual Simultaneous Localisation and Mapping (VSLAM) in order to use the camera on the front of the drone to localise and map the environment, from this it should be possible to determine distance from obstacles and travel through the hall way, however the doorframe may be another issue all together! I may post more information about what we do as the project goes on.

The project so far

As term has only just begun, so has the project. At the moment we are working to complete the project initiation documentation (PID) by this coming Friday. After this is complete, we shall start working with the drone and looking forward for the project.

Third year lectures

I have now had a lecture from each of the modules for this term and some certainly already seem that they are going to be much harder than others. This is something I expected however as I do a mix of modules from IT, cybernetics and robotics. One of the most interesting modules I am currently doing is Machines in Motion, a module that looks at the kinematics of robotic manipulators in open chain systems. This module is also the hardest module I have!

Lectures this year have much smaller class sizes on the whole, which means more lecturer interaction. This is good as it allows the class to feedback to the lecturer on things that require more time spent on them.

Although the modules this year are harder than last year on the whole, most have open-book exams. This means textbooks and notes are allowed to be taken into the exam venue with you. Unfortunately this does not necessarily make thing easier, as there is not time in the exam to look up the answer to every question!