2IOE0 Interactive Intelligent Systems
Erik de Vink, Huub van de Wetering
TU/e
2014-2015
Erik de Vink, Huub van de Wetering (TU/e)
2IOE0 Interactive Intelligent Systems
2014-2015
1 / 18
Introduction
Course resources
1
http://www.win.tue.nl/˜wstahw/edu/2IOE0
2
http://www.win.tue.nl/˜wstahw/edu/2IOE0/studyguide.html
2IOE0 = Two India Oscar Echo Zero
Erik de Vink, Huub van de Wetering (TU/e)
2IOE0 Interactive Intelligent Systems
2014-2015
2 / 18
Introduction
Design Based Learning
1
Work in a team of upto 6 participants.
2
Apply knowledge of previous courses.
2IV60 Computer Graphics
2ID90 Artificial Intelligence
3
Process guided by tutors.
Erik de Vink
Huub van de Wetering
4
Realize an interactive intelligent system.
2014/15: a game
Erik de Vink, Huub van de Wetering (TU/e)
2IOE0 Interactive Intelligent Systems
2014-2015
3 / 18
Introduction
Team composition
Time slots
E (obligatory).
either A (Monday morning, Thursday afternoon)
or B (Monday afternoon, Wednesday morning)
In the break
find your favourite team mates
with matching time slot preference.
Restrictions
Team size either 5 or 6.
|#teams in A − #teams in B| ≤ 1
Remaining issues are solved by the tutors.
Erik de Vink, Huub van de Wetering (TU/e)
2IOE0 Interactive Intelligent Systems
2014-2015
4 / 18
Introduction
Team resources
Each team has a project room
will be announced
Each team is monitored by a tutor
time slot B (Monday afternoon, Wednesday morning)
Erik de Vink
time slot A (Monday morning, Thursday afternoon)
Huub van de Wetering
Erik de Vink, Huub van de Wetering (TU/e)
2IOE0 Interactive Intelligent Systems
2014-2015
5 / 18
Introduction
Project execution
For the rest, read study guide!!!
Erik de Vink, Huub van de Wetering (TU/e)
2IOE0 Interactive Intelligent Systems
2014-2015
6 / 18
Interactive Intelligent System: Game
should contain challenging/non-trivial/interesting elements of
Computer Graphics
geometry, rendering, algorithms, efficiency, HCI, ...
Artificial Intelligence
searching, planning, learning, ...
Supporting tools may also earn you credits
if properly designed, implemented, demonstrated, and reported upon.
Erik de Vink, Huub van de Wetering (TU/e)
2IOE0 Interactive Intelligent Systems
2014-2015
7 / 18
Terminology
PC or player character:
character controlled by a human player
e.g. pac-man
NPC or non-player character:
the other characters
e.g. the ghosts in Pac-Man
Erik de Vink, Huub van de Wetering (TU/e)
2IOE0 Interactive Intelligent Systems
2014-2015
8 / 18
General Game Architecture
Game state:
contains current state of the world
may be queried on state of objects in the world
Erik de Vink, Huub van de Wetering (TU/e)
2IOE0 Interactive Intelligent Systems
2014-2015
9 / 18
General Game Architecture
Controller:
each character has at least one
selects actions
PC: controller interprets “joy stick” events
NPC: controller is character’s brain: AI, ...
Erik de Vink, Huub van de Wetering (TU/e)
2IOE0 Interactive Intelligent Systems
2014-2015
9 / 18
General Game Architecture
Renderer:
renders the game state
outputs images (and sounds)
Erik de Vink, Huub van de Wetering (TU/e)
2IOE0 Interactive Intelligent Systems
2014-2015
9 / 18
General Game Architecture
Simulator:
encodes how game state changes
responds actions of the controllers
Erik de Vink, Huub van de Wetering (TU/e)
2IOE0 Interactive Intelligent Systems
2014-2015
9 / 18
General Game Architecture
Erik de Vink, Huub van de Wetering (TU/e)
2IOE0 Interactive Intelligent Systems
2014-2015
10 / 18
Controllers
either programmed or learned
reactive controller:
stateless
function of its input
stochastic controller
more interesting to play against non-predictable NPC
randomness may break symmetry to avoid, e.g. being trapped in a
corner
mimicks noise in real world
for instance, add randomness to controller’s output
Erik de Vink, Huub van de Wetering (TU/e)
2IOE0 Interactive Intelligent Systems
2014-2015
11 / 18
Controllers
Hierararchy
several controllers together may determine the next (game) action,
e.g.
one controller determines goal of NPC
a second controller searches a path to the goal
a third controller computes next move.
Erik de Vink, Huub van de Wetering (TU/e)
2IOE0 Interactive Intelligent Systems
2014-2015
12 / 18
Controllers
Percept
A percept
functions as a filter/facade for the game state.
game state is only available via its methods.
may be replaced by another percept, for example,
one which uses other info from gamestate,
a ’confused’ one with randomized values if
(N)PC has taken a hit on the head.
Erik de Vink, Huub van de Wetering (TU/e)
2IOE0 Interactive Intelligent Systems
2014-2015
13 / 18
Game Design
Players: how many? Any requirements? Special knowledge, roles ... ?
Objective: What is the objective of the game?
Procedures: What are the required actions?
Rules: Any limits on player actions? Rules regarding behaviour?
Conflict: What causes conflicts?
Outcome: What are the possible outcomes?
...
Erik de Vink, Huub van de Wetering (TU/e)
2IOE0 Interactive Intelligent Systems
2014-2015
14 / 18
Game Design
All video game types are allowed:
action, adventure, role-playing, simulation, strategy, sports, ....
But,be realistic
not everything is possible in one quartile.
Make the CG and AI shine!
Erik de Vink, Huub van de Wetering (TU/e)
2IOE0 Interactive Intelligent Systems
2014-2015
15 / 18
Game Realization
Build modular
decide on interfaces
document interfaces
Use fast prototyping
iteratively build your game
Play testing
find problems
tune parameters
Erik de Vink, Huub van de Wetering (TU/e)
2IOE0 Interactive Intelligent Systems
2014-2015
16 / 18
First Team Actions
1
Make weekly appointment with tutor.
2
Read the studyguide.
3
Brainstorm on game concepts.
4
Work out a concept on paper.
5
...
Erik de Vink, Huub van de Wetering (TU/e)
2IOE0 Interactive Intelligent Systems
2014-2015
17 / 18
References
1
Artificial Intelligence for Computer Graphics, John David Funge.
2
Game Design Workshop, Tracy Fullerton.
Erik de Vink, Huub van de Wetering (TU/e)
2IOE0 Interactive Intelligent Systems
2014-2015
18 / 18