The project presentation slides in Microstoft ... - MyreSpace

almondpitterpatterAI and Robotics

Feb 23, 2014 (3 years and 1 month ago)

68 views

Enemy Agent

Responding to stimuli in a real time
3D environment

The Problem



To create a
a real
-
time, AI agent

in PC
computer game
‘Half
-
Life’

that can
respond
to a wide array of stimuli in a fast paced
environment

and be
an enjoyable challenge
to fight in the game.


Designing the Agent


The concept of a game ‘modification’ and its
relation to all areas of modern coding practice.


Other research in this area. [2]


Modelling behaviours on territorial animals,
leading to a focus on non
-
typical computer game
behaviours. [5]


Designing the ‘fun factor’; immersion and
evoking
emotions in a human game player.[1]

The Plan


Basic functionality first style of design.[3]


An agent that could respond interestingly to
a wide variety of player behaviours rather
than just blindly attacking.


Putting the intelligence into the
environment [4].


Working together with other agents.

Enemy Agent Class Diagram

Red = Completely original to this project

Green = Modified for this project

Blue = Standard
Half
-
Life

class

The Implementation


Using a new language; C++


The challenges and advantages of working with
someone else’s code.


3D aspect required relearning of mathematics of
trigonometry and vectors.


Extensive non coding materials required to get
project functioning.

Agent’s in Action

0
2
4
6
8
10
12
0
1
2
3
4
5
6
7
Hunger
Fear
Anger

Three key
variables
control the
agents
behaviour


What the project achieved


Created a functioning and hopefully fun in
-
game opponent.


Made the in
-
game opponent display a
variety of interesting responses to player
driven stimuli.


Successfully overcame implementation
difficulties to implement the crucial
specification requirements.


Evaluation


Some fringe aspects of the specification not
completed due to difficulty and time constraints.


Learned a huge array of new skills and feel the
project was well planned and very satisfactorily
delivered.


Future work could focus on agents developing
learning behaviour and reacting to a wider range
of emotional drives, increasing the believability of
the agents behaviour.

Summary


Project achieved its key aims.


Targeted and background research has
achieved excellent understanding of the
subject area.


Agent was well planned and designed in a
very robust way.

References

[1]

Agents that Want and Like: Motivational and Emotional Roots of Cognition and Action
-

Papers from the AISB’05 Symposium

by

SSAISB Press


[2]
It knows What You Are Going To Do: Adding anticipation to a Quakebot


By

John E.
Laird


Published
2001

http://ai.eecs.umich.edu/people/laird/papers/Agents01.pdf
available on 21
st

April 2005


[3]
Game Architecture and Design: A New Edition

-

Andrew Rollings and Dave Morris

-

Published

2003 by New Riders


[4] Solving the right problem


Neil Kirby, Bell Laboratories



Published 2002 AI Game
Programming Wisdom, Charles River Media


[5] Territorial intrusion risk and antipredator behaviour: a mathematical model
-

Diaz
-
Uriarte R
-

Proceedings: Biological Sciences, Volume 268, Number 1472 / June 7, 2001