School of IT

ICT219 Intelligent Systems
Semester 2, 2007

ICT219 Intelligent Systems Unit Information

Unit Coordinator and Lecturer

Consultation Times

Time	Room
Mondays 2:30-3:30	ECL2.061
Fridays 11:30-12:30	ECL2.061

Tutor (including External Students)

Roger Wright

Textbook

Champandard, Alex. AI Game Development. New Riders Publishers, 2003.

If you miss out on a copy from the bookshop get it from Amazon.com

Notice Board

(Please read this notice board regularly for the latest news.)
(Click on "Refresh" or "Reload" to read the latest notices)

19-11-07. Good luck to everyone with your exams. If you need help with preparing, visit me during the times mentioned in the final lecture.

5-11-07. In this final week, I’d like everyone to focus on getting all your submission finished and submitted. Even if you’re late with these, still put them in - every extra mark counts! And don’t foget this week’s lecture full of hints, tips and tricks for the final exam.

26-10-07. Details on how to submit your project are in the Unit Information handout. Don’t forget your to include a signed cover sheet . ~o~ The credit for those Small coordinate transformation functions you might be using in your Guntactyx code should go to Jeff Timlin, a fellow student. Make sure you acknowledge that if you use those.

19-10-07. I’ve added a link below to the Small manual. This is a short version, which should answer most questions, but if you want the full story, there’s a link in it to the “Small booklet” which isn’t small at all - it’s a 155-page doorstopper.

15-10-07. Some folks didn’t put in a signed cover sheet with their Labs 1-6. If you didn’t, please submit one separately, either on paper or by email to your tutor - we need that to give you the marks. ~o~ As I mentioned, our labs are a little out of order this semester. This week’s concerns a rule-based system, which we covered in lectures a couple of weeks ago. At least you should be across the theory for this by now.

24-9-07. For the theory behind this week’s lab, you need to look ahead to the lecture for Week 8, concerning neural networks. We are a little out of sequence with our labs, and we still need to cover the remaining material from Week 6 on rule based systems. For this reason, we’ll catch up by skipping straight to Topic 8 after the break.

24-9-07. Sorry for those of you who were inconvenienced in the labs last week. Roger Wright was called away suddenly on account of a death in his family, and I didn’t discover this until late Friday because I was at a conference for most of the week (otherwise I would have covered his classes). I hope you were able to carry on with that methodology exercise anyway. It is, of course, mostly a way of getting you thinking systematically about the design of your project sooner, rather than later.

10-9-07. I’d like to remind everyone to keep up with your reading from the textbook, as well as work on your practical lab exercises. Remember, we want to see properly written-up answers to the questions on the bottom of Labs 1-7 by September 28th. Even if you can’t make it to the weekly labs, still work on those on your own machines, each week if possible.

This week, we are working on understanding an emotional animat called Moody which uses the FSA and FSM methods we used last week. Before you try Lab 4, it will be worth reading through this week’s lecture and Chapters 37, 39 and 42 of Champandard, which put forward a model of how emotions might work in an animat. It will be difficult to understand the Moody code without this background knowledge. Also, there’s a slightly improved version of Moody on the website (under Download Software, below).

20-8-07. The project spec is released - see below for details. It is a choice between a FEAR/Quake 2 animat or a Guntactyx team of soccer robots. Have a look at that and make a start as soon as you have chosen one - there’s precious little time this semester.

20-8-07. This week we’ll study the theory and practice of reactive design as it applies to animats and bots. We started talking about that in last week’s lecture, you’ll recall, and I’ll say more about it in this week’s. The lab requires you to read and understand the Small code for the simple reactive team bot called recruit.
To understand that, you’ll need to read through the Guntactyx manual, which is in the Guntactyx directory. I want you to learn how to modify and compile your own .sma files and try to get them running in the game. Hint: Press the F3 key to bring up a transparent blue “heads-up” display screen, which lets you control editing, compiling and executing new bot scripts.

15-8-07. Okay, the FEAR/Quake2 software for the labs has tested out okay, and should be available in the labs for Friday. You should download Lab1 below and have a look at that. Remember to write up your answers to the questions as they will need to be submitted to your tutor during the semester.

13-8-07. Well, the FEAR/Q2 platform still hasn’t been installed into the labs, but I hope I can have it installed there by the end of the week. The other software is there, though. And there’s no reason you couldn’t install the FEAR/Quake2 platform on a home machine. I don’t have enough disks for everyone, but I will try to circulate as many copies as possible, if you could just pass them on to the next student when you are done.

8-8-07. Hi there! I hope you enjoy studying artificial intelligence by developing and playing computer games. I’ve still got a some work to do to get everything functioning perfectly. The lecture slides and labs will be slightly updated as we go along, so I won’t release these all at once. Please be patient if some software or document is not available (but phone or email me if you spot some problem). To get ready for the unit make sure you have the textbook and some removable media (CD-R disks).

Lecture Schedule

Time	Location	Lecturer
Fridays 2:30-4:30pm	ECL3	Graham Mann

(These could change as we go through the semester, so download them as late as you can.)

Lecture Slides	Title	Lab Exercises
Week 1	Games, AI and AI in Games; A Brief History of AI	None
Week 2	Reactive and Reflective Agents; Animats; Game Development and Game Platforms	Lab 1
Week 3	Obstacle avoidance; A Reactive animat, Problem Solving	Lab 2
Week 4	State Machines; Control of an Animat by States	Lab 3
Week 5	Emotive States; an Emotive Animat	Lab 4
Week 6	Knowledge Representation, Rule Based Systems; A Rule-Based Animat	Lab 5
Week 7	The Project; AI Methodology and Evaluation	Lab 6
Week 8	Machine Learning; Neural Networks	Lab 7
Week9	Neural Networks; A Neural Network Animat	Lab 8
Week10	Genetic Algorithms	Lab 9
Week11	A Genetic Animat	Lab 10
Week12	Revision and Exam Preparation	Finishing up remaining work

Project

Project Specification

Cover Sheet (Off-campus)

Cover Sheet (On-campus)

Tutorials

Time	Laboratory	Tutor
Mon 12:30-1:30	ECL2.005	Graham Mann
Fri 12:30-1:30	ECL2.005	Roger Wright
Fri 1:30-2:30	ECL2.005	Roger Wright
Fri 4:30-5:30	ECL2.005	Roger Wright