[PPT] - CS 480/680: GAME ENGINE PROGRAMMING ARTIFICIAL INTELLIGENCE PowerPoint Presentation

SLIDE 1

CS 480/680: GAME ENGINE PROGRAMMING ARTIFICIAL INTELLIGENCE

3/7/2013 Santiago Ontañón santi@cs.drexel.edu https://www.cs.drexel.edu/~santi/teaching/2013/CS480-680/intro.html

SLIDE 2

Game AI in Educational Game

A CS RA @ 20 hrs/week for two-terms in the Digital Media program. Experience Management for educational games Skills: * basic knowledge in AI * C# and/or JavaScript * Unity 3D experience is a plus If you are interested, contact me.

SLIDE 3

Outline

Student Presentations
Game AI
Movement
Path-finding
Scripting
Board Games
Learning
Project Discussion

SLIDE 4

Outline

Student Presentations
Game AI
Movement
Path-finding
Scripting
Board Games
Learning
Project Discussion

SLIDE 5

Student Presentations

Michael Parenzan:
“Game Audio Design Patterns”
William Stern:
“Classic Super Mario 64 Third-Person Control and Animation”
Ryan Daugherty:
“Time Bounded A*: Efficient Real Time Pathfinding”
Andrew Townsley
“Time and Consistency Management for Multiserver-Based

MMORPGs”

Thomas Burdak:
“Vector and Plane Tricks”

SLIDE 6

Outline

Student Presentations
Game AI
Movement
Path-finding
Scripting
Board Games
Learning
Project Discussion

SLIDE 7

Game Engine Architecture

HARDWARE DRIVERS OS SDKs Platform Independence Layer Utility Layer Resource Management Game Engine Functionalities Game Specific Game Engine Dependencies

SLIDE 8

Game Engine Architecture

Rendering Engine Animation Engine Collisions Physics Audio Subsystem Online Multiplayer Profiling & Debugging Gameplay Foundations (Game Loop) Artificial Intelligence Scripting

SLIDE 9

Game Engine Architecture

Rendering Engine Animation Engine Collisions Physics Audio Subsystem Online Multiplayer Profiling & Debugging Gameplay Foundations (Game Loop) Artificial Intelligence Scripting

SLIDE 10

What is Game AI?

Artificial Intelligence for Computer Games
Different from traditional AI

Traditional AI: Optimality, efficiency Game AI: Fun, artificial “stupidity”

SLIDE 11

Examples of Game AI

Pac-Man (1980) First ever game to feature AI AI: finite state machine

SLIDE 12

Examples of Game AI

“Pac-Man” (1980) First ever game to feature AI AI: finite state machine

SLIDE 13

Examples of Game AI

“Double Dragon” AI: Finite State Machines

SLIDE 14

Examples of Game AI

Chess AI needs to provide a collection of difficulty levels. Only one: hardest (to be played only against grand-masters), falls into the realm of traditional AI.

SLIDE 15

Examples of Game AI

“Left 4 Dead 2” AI Director adjusts game pace to ensure desired dramatic effects

SLIDE 16

Examples of Game AI

“Starcraft II” Strategy, planning, pathfinding, economics, etc.

SLIDE 17

AI Interface with the Game Engine

Game State Collision Input AI World Interface (perception) Strategy Decision Making Movement

SLIDE 18

Example: RPG Game

Consider Morrowind (Elder Scrolls III)

SLIDE 19

Example: RPG Game

Game State Collision + Line of sight Input AI Perception:

Navigation

mesh

Quests’

status

Interaction

events (hits, etc.) Strategy Decision Making: Scripts Movement: A* + Animation

SLIDE 20

Example: RPG Game

Game State Collision + Line of sight Input AI Perception:

Navigation

mesh

Quests’

status

Interaction

events (hits, etc.) Strategy Decision Making: Scripts Movement: A* + Animation For the triggers and conditions used in the scripts (basically, we need all the information needed to test the conditions in the scripts, or an interface to do so)

SLIDE 21

Example: RPG Game

Game State Collision + Line of sight Input AI Perception:

Navigation

mesh

Quests’

status

Interaction

events (hits, etc.) Strategy Decision Making: Scripts Movement: A* + Animation Basically for A*, although some scripts might check for things like “if there is path between A and B then …”

SLIDE 22

Polling

Simplest idea:
No special perception layer
Whenever the AI wants some piece of information, it just asks the

Game State / Collision / Line of Sight modules about it

Benefits:
Easy
Problems:
Inefficient
The same information might be asked for multiple times
Most of the time the player does not collide with enemies, why

check it at every frame?

Hard to debug

SLIDE 23

Message Passing

Event-based approach
Wait for an event message to arrive rather than constantly

checking for conditions to be true at each cycle (e.g. check if the player is in line of sight):

When something relevant happens (e.g. player is in line of sight), a

message is sent to all the AIs that care about this condition.

Centralized “event manager”

SLIDE 24

Event Manager

Checking Engine Event queue Registry

f

Event Listeners AI1 AI2 AIn

SLIDE 25

Event Manager

Checking Engine Event queue Registry

f

Event Listeners AI1 AI2 AIn AIs register for certain events, like collisions with certain objects, etc.

SLIDE 26

Event Manager

Checking Engine Event queue Registry

f

Event Listeners AI1 AI2 AIn A single checking engine checks for the events for which any AI is registered

SLIDE 27

Event Manager

Checking Engine Event queue Registry

f

Event Listeners AI1 AI2 AIn Certain events, do not need to be checked for, since the originating entity sends the directly (e.g. state changes in another AI)

SLIDE 28

Outline

Student Presentations
Game AI
Movement
Path-finding
Scripting
Board Games
Learning
Project Discussion

SLIDE 29

Game AI Architecture

AI World Interface (perception) Strategy Decision Making Movement

SLIDE 30

Steering Behaviors

Basic building blocks for continuous movement
Whole family of steering behaviors
Widespread use among commercial computer games

SLIDE 31

Steering Behaviors: Uses

SLIDE 32

Steering Behaviors: Uses

Decision Making Movement

?

SLIDE 33

Steering Behaviors: Uses

Decision Making Movement In car racing games, Decision Making is typically hard coded. The game designers create a set

f waypoints in the track (or in the

track pieces), and cars go to them in order.

SLIDE 34

Steering Behaviors: Uses

Decision Making Movement Movement is in charge of driving the car to each of the waypoints, avoiding opponents, braking, accelerating, turning, etc.

SLIDE 35

Steering Behaviors: Uses

Not just racing games
Any games with vehicles (helicopters, tanks, planes,

boats)

Or even characters moving in a 3D environment

(continuous movement) with inertia (e.g. sports games)

Most FPS games just assume there is no inertia and characters

can move in any direction at any time (bad physics J)

SLIDE 36

Basic Steering Behaviors

Seek
Flee
Arrive
Align
Velocity Matching

SLIDE 37

Steering Behaviors

Defined as methods that return the acceleration that the

body/vehicle controller by the AI needs to have during the next execution frame:

Input: position, orientation, speed of AI, target
Output: acceleration
They are executed once per game cycle
Some return linear acceleration (accelerate north at 3m/

s2), some return angular acceleration (turn right at 2rad/s2)

SLIDE 38

Seek

Move towards a (potentially moving) target

Target

SLIDE 39

Seek

Move towards a (potentially moving) target

Target Difference D = E - S S: Start coordinates E: End coordinates V: Current Speed

SLIDE 40

Seek

Seek(character, E) D = E - character.position ND = D / |D| A = ND * maxAcceleration Return A

SLIDE 41

Seek

Seek(character, E) D = E - character.position ND = D / |D| A = ND * maxAcceleration Return A

Target character.position E: End coordinates character.velocity

SLIDE 42

Seek

Seek(character, E) D = E - character.position ND = D / |D| A = ND * maxAcceleration Return A

Target character.position E: End coordinates character.velocity Difference D = E – character.position

SLIDE 43

Seek

Seek(character, E) D = E - character.position ND = D / |D| A = ND * maxAcceleration Return A

Target character.position E: End coordinates character.velocity ND

SLIDE 44

Seek

Seek(character, E) D = E - character.position ND = D / |D| A = ND * maxAcceleration Return A

Target character.position E: End coordinates character.velocity A

SLIDE 45

Seek

Seek(character, E) D = E - character.position ND = D / |D| A = ND * maxAcceleration Return A

Target character.position E: End coordinates character.velocity A

SLIDE 46

Seek

Seek(character, E) D = E - character.position ND = D / |D| A = ND * maxAcceleration Return A

Target character.position E: End coordinates character.velocity A

SLIDE 47

Seek

Seek(character, E) D = E - character.position ND = D / |D| A = ND * maxAcceleration Return A

Target character.position E: End coordinates character.velocity A

SLIDE 48

Composite Steering Behaviors

Pursue and Evade
Face
Looking where you are going
Wander
Path Following
Separation
Collision Avoidance
Obstacle/Wall Avoidance
General Combination

SLIDE 49

Steering Behaviors in Vehicles

As defined so far, steering behaviors assume:
Character/vehicle under control can exert a force at an arbitrary

angle

The direction of movement is independent of the direction being

faced

None of those assumptions are satisfied by vehicles
Are Steering behaviors still useful then?

SLIDE 50

Motor Control Layer

Steering Behaviors generate the “desired accelerations”.

An underlying “motor layer” translates that into commands like “accelerate, brake, turn right, turn left”:

AI World Interface (perception) Strategy Decision Making Movement Steering Behaviors Motor Control

SLIDE 51

Output Filtering

Idea:
Use the steering behavior to produce an acceleration request
Project the request onto the accelerations that the vehicle at hand

can perform, and ignore the rest

Steering Request: Vehicle capabilities: Projection:

SLIDE 52

Outline

Student Presentations
Game AI
Movement
Path-finding
Scripting
Board Games
Learning
Project Discussion

SLIDE 53

Pathfinding

Problem:
Finding a path for a

character/unit to move from point A to point B

One of the most common

AI requirements in games

Specially critical in RTS

games since there are lots of units

A B

SLIDE 54

Pathfinding

Problem:
Finding a path for a

character/unit to move from point A to point B

One of the most common

AI requirements in games

Specially critical in RTS

games since there are lots of units

A B

SLIDE 55

Pathfinding

Simplest scenario:
Single character
Non-real time
Grid
Static world
Solution: A*
Complex scenario:
Multiple characters (overlapping paths)
Real time
Continuous map
Dynamic world

SLIDE 56

Pathfinding is a Problem!

Even in modern commercial games:
http://www.youtube.com/watch?

v=lw9G-8gL5o0&feature=player_embedded

SLIDE 57

Quantization and Localization

Pathfinding computations are performed with an

abstraction over the game map (a graph, composed of nodes and links)

Quantization:
Game map coordinates -> graph node
Localization:
Graph node -> Game map coordinates

SLIDE 58

Tile Graphs

Divide the game map in equal tiles (squares, hexagons,

etc.)

Typical in RTS games

SLIDE 59

Navigation Meshes (Navmesh)

By far the most widely used
Use the level geometry as the pathfinding graph
Floor is made out of triangles:
Use the center of each “floor” triangle as a characteristic point

SLIDE 60

Navigation Meshes (Navmesh)

One characteristic point is only connected to neighboring
nes

SLIDE 61

Navigation Meshes (Navmesh)

The validity of the graph depends on the level designer. It

is her responsibility to author proper triangles for navigation:

SLIDE 62

Pathfinding

Simplest scenario:
Single character
Non-real time
Grid
Static world
Solution: A*
Complex scenario:
Multiple characters/Dynamic world: D* / D* Lite
Real time: TBA* / LRTA*
Continuous map: Quantization

SLIDE 63

Outline

Student Presentations
Game AI
Movement
Path-finding
Scripting
Board Games
Learning
Project Discussion

SLIDE 64

Scripting

Most Game AI is scripted
Game Engine allows game developers to specify the

behavior of characters in some scripting language:

Lua
Scheme
Python
Etc.
Game specific languages

SLIDE 65

Scripting

Advantages:
Gives control to the game designers (characters do whatever the

game designers want)

Disadvantages:
Labor intensive: the behavior of each characters needs to be

predefined for all situations

Rigid:
If the player finds a hole in one of the behaviors, she can exploit it again

and again

Not adaptive to what the player does (unless scripted coded to be so)

SLIDE 66

Finite State Machines

Harvest Minerals Build Barracks Train marines Attack Enemy Explore Train SCVs 4 SCVs harvesting 4 SCVs Less than 4 SCVs barracks 4 marines & Enemy seen Enemy seen 4 marines & Enemy unseen No marines

SLIDE 67

Finite State Machines

Easy to implement:

switch(state) { case START: if (numSCVs<4) state = TRAIN_SCVs; if (numHarvestingSCVs>=4) state = BUILD_BARRACKS; Unit *SCV = findIdleSCV(); Unit *mineral = findClosestMineral(SCV); SCV->harvest(mineral); break; case TRAIN_SCVs: if (numSCVs>=4) state = START; Unit *base = findIdleBase(); base->train(UnitType::SCV); break; case BUILD_BARRACKS: … }

SLIDE 68

Finite State Machines

Good for simple AIs
Become unmanageable for complex tasks
Hard to maintain

SLIDE 69

Finite State Machines (Add a new state)

Harvest Minerals Build Barracks Train marines Attack Enemy Explore Train SCVs 4 SCVs harvesting 4 SCVs Less than 4 SCVs barracks 4 marines & Enemy seen Enemy seen 4 marines & Enemy unseen No marines

?

SLIDE 70

Finite State Machines (Add a new state)

Harvest Minerals Build Barracks Train 4 marines Attack Enemy Explore Train 4 SCVs 4 SCVs harvesting 4 SCVs Less than 4 SCVs barracks 4 marines & Enemy seen Enemy seen Enemy unseen No marines Attack Inside Enemy Enemy Inside Base

SLIDE 71

Standard Strategy

Hierarchical Finite State Machines

Attack Inside Enemy Enemy Inside Base No Enemy Inside Base

FSM inside of the state of another FSM
As many levels as needed
Can alleviate complexity problem to some extent

SLIDE 72

Hierarchical Finite State Machines

Harvest Minerals Build Barracks Train 4 marines Attack Enemy Explore Train 4 SCVs 4 SCVs harvesting 4 SCVs Less than 4 SCVs barracks 4 marines & Enemy seen Enemy seen Enemy unseen No marines

Attack Inside Enemy Enemy Inside Base No Enemy Inside Base

FSM inside of the state of another FSM
As many levels as needed
Can alleviate complexity problem to some extent

SLIDE 73

Behavior Trees

Combination of techniques (some of them we covered last

week):

Hierarchical state machines
Scheduling
Automated planning
Action Execution
Increasingly popular in commercial games
Strength:
Visual and easy to understand way to author behaviors and

decisions for characters without having programming knowledge

SLIDE 74

Behavior Trees Appeared in Halo 2

Halo 2 (2004)

SLIDE 75

Example Behavior Tree

SLIDE 76

Behavior Tree Basics

A behavior tree (BT) captures the “behavior” or “decision

mechanism” of a character in a game

At each frame (if synchronous):
The game engine executes one cycle of the BT:
As a side effect of execution, a BT executes actions (that control the character)
The basic component of a behavior tree is a task
At each game cycle, a cycle of a task is executed
It returns success, failure, error, etc.
As a side effect of execution they might execute things in the game
Three basic types of tasks:
Conditions
Actions
Composites

SLIDE 77

Behavior Tree Tasks

Action Condition Sequence Selector Generic: the same for all games Domain dependent: Each game defines its own actions and conditions

SLIDE 78