This paper’s study is the development of the mobile phone game based on Android OS. In this paper, we present the design and implementation of an Android game we will develop, called Sammy. In this game the player will select between available tiles to create a path for Sammy to reach to his home from school. Sammy is developed for mobile devices with the Android mobile operating system. The tool used for developing this game is Unity 3D game engine. The game will also have an A.I. player which the player can play against. Nowadays mobile gaming market is growing rapidly and is expected to be a $16 billion business by 2016. Almost all Android devices will be able to successfully run the game. We hope that sharing our experiences will assist others who wish to either use our mobile game or develop their own. Our research seeks to adding values and enjoyment to users of the Android mobile devices.
1
3D ANDRIOD GAME DEVELOPMENT
WITH A.I
Vrushab Randive Shubham Vernekar, Chandrasen Kumakar, Monika Shirbhate
Developer,Developer,Developer,Developer
Computer Department
G.H. RAISONI INSTITUTE OF ENGINEERING AND TECHNOLGY,PUNE,INDIA
Abstract -- this paper's study is the development of the
mobile phone game based on Android OS. In this paper,
we present the design and implementation of an
Android game we will develop, called Sammy. In this
game the player will select between available tiles to
create a path for Sammy to reach to his home from
school. Sammy is developed for mobile devices with the
Android mobile operating system. The tool used for
developing this game is Unity 3D game engine. The
game will also have an A.I. player which the player can
play against. Nowadays mobile gaming market is
growing rapidly and is expected to be a $16 billion
business by 2016. Almost all Android devices will be
able to successfully run the game. We hope that sharing
our experiences will assist others who wish to either use
our mobile game or develop their own. Our research
seeks to adding values and enjoyment to users of the
Android mobile devices. Key Words: Game Engines;
Animation; Scripting; Android;
I. Introduction
The game Sammy is developed for mobile devices with an
Android operating system, which was designed as an
operating system for mobile devices by Google and the
Open Handset Alliance based on the Linux kernel. The
game Sammy is a strategy game that lets player select and
place a number of tiles to form a path from Sammy's
school to his house. The game has two game modes
(Predefined levels and VS. A.I.) And three levels of
difficulty (easy, intermediate, and hard). Each tile will have
its own unique specialties and the player has to strategize to
find the best path for maximum score. Score will be
awarded on the basis of time spent to complete the level,
the number of tiles used and collecting stars. Each level
will have 3 stars which will give bonus to the player.
Casual gamers tend to enjoy simple, yet dynamic
games that are easy to understand, frequently reward the
player, are short in duration (as opposed to games where
characters must be developed and nurtured), and have high
re-playability by not becoming boring or repetitive. Our
game achieves all these goals by adding new game play
features and tiles as the level progresses.
1.1
Product Features
A high level list of the major features of our
developed game is as follows.
·
Two operational modes: single player and Vs.
A.I.
·
Three tiers of difficulty: easy, intermediate, and
hard
·
Invalid placements are not allowed.
·
In-game rule-set, tutorial mode where the rules of
the game are explained.
·
Intuitive and linear interface.
·
Smooth animation system.
·
Bright, vibrant colors.
1.2
User Interface
Usability and portability are the most important quality
attributes applicable to this research project. Since the
game Sammy is targeted toward casual gamers, the
interface and mechanics of game play are required to be
simple and straight forward. Upon initial launch, the game
enters the main screen, where the user can select one of the
game modes (Single player and Vs. A.I.) or mess with the
settings such as change volume or change difficulty. After
game mode the system has enough information to initialize
the game and allow game play to begin. Game play State is
not a single state but an aggregation of several states like
selection of tiles, placing of tiles, scoring. Upon completion
of game play, the system transitions to the End Game State,
where the user is notified of his success or failure. When
the user does an invalid placement of tiles he will be
notified of invalid movements, he will see a text message
appear in a notification area after an invalid placement
attempt.
2
II. Literature Survey
·
Study of Unity 3D game engine for use of its
physics and texture engines
·
Use of MAYA and Adobe Photoshop for model
and background design
·
Nowadays many big companies have developed
huge and interesting android games that have set
the bar high
·
In this independent game with our limited
resources we will attempt to recreate the same
spark as games developed by these big games.
III. Game engines
A game engine is a software framework designed for the
creation and development of video games.
The core functionality typically provided by a game engine
includes
1.
Rendering engine ("renderer") for 2D or 3D
graphics
2.
Physics engine or collision detection
3.
Animation
4.
Scripting
5.
A. I. and networking and much more
Fig 1. Logo of Frostbite Game Engine
[image courtesy/source:
http://wegotthiscovered.com/gaming/ea-bringing-frostbite-
engine-mobile-platforms-skipping-wii/]
4.1. Unity - Game Engine
Unity is a game development ecosystem: a powerful
rendering engine fully integrated with a complete set of
intuitive tools and rapid workflows to create interactive 3D
and 2D content.
Unity provides all the tools a game developer need to
develop a 2D or 3D game on multiple platforms with
minimum cost and time which makes it ideal for
independent developers like us.
Fig 2. Logo of Unity Game Engine
[image courtesy/source:
http://imgkid.com/unity-3d-logo-png.shtml]
IV. Android Operating System
Android is a comprehensive operating environment that
based on Linux® V2.6 kernel, it is also a layered system,
the architecture of Android system is shown in the picture
Fig 3. Architecture of Android
[image courtesy/source:
http://www.techotopia.com/index.php/An_Overview_of_th
e_Kindle_Fire_Android_Architecture]
Applications layer is the site of all Android applications
including an email client, SMS program, maps, browser,
contacts, and others. All applications are written using the
Java programming language. Application framework layer
defined the Android application framework. All Android
applications are based on the application framework. The
Android application framework including:
· A rich and extensible set of Views that can be used to
build an application with beautiful user interface, including
lists, grids, text boxes, buttons, and even an embeddable
web browser.
· A set of Content Providers that enable applications to
access data from other applications (such as Contacts), or to
share their own data.
· A Resource Manager that provides access to non-code
resources such as localized strings, graphics, and layout
files.
· A Notification Manager that enables all applications to
display custom alerts in the status bar.
· An Activity Manager that manages the lifecycle of
applications and provides a common navigation back stack.
Libraries layer includes a set of C/C++ libraries used by
various components of the Android system and provides
support to the application framework. [2]
3
V. Algorithms
These are the algorithms used in our project for the two
game modes and also the behavior of A.I.
6.1. Algorithm
1.
The game will start with a GUI allowing player to
1.
Compete against A.I. opponent
2.
Try to beat predefined levels and set high
scores
3.
Mess with game setting such as sound on or
off
2.
When a new game is started the game will generate a
level which will be a n*m size board (where n = no
of rows and m = no of columns) made of square tiles.
3.
Each of these levels will have a predefined starting
point tile and an ending point tile
4.
Each level will also have 3 stars on tiles which when
obtained will give special points
5.
The player will be given a set of tiles to choose from
and objective of the player is to take the object from
the starting point to the finish point in least time and
most score
6.
Each tile will have special properties and it's the
players task to choose the best combination of these
tiles to get the most score
7.
After player has decided the path the game will show
the simulation of the object going from the start point
to the end
8.
At the end of level player will be shown various
statistics about his/her performance.
9.
The player will be advanced to next level and a new
level will be generated
6.2. Algorithm: Predefined Levels
Here each level will have a starting tile and a finish tile.
1.
Each level will also feature some predefined tiles that
must be used by the player and 3 stars which will
grant bonus score.
2.
The objective of the player is to take the object from
start tile to end tile and try to make the highest score
3.
The player will be allowed to choose from a set of
tiles to form a path to take the object from start to
finish
4.
Once all tiles are selected the object will be released
from start tile and pass through the various tiles
5.
Each tile will have its own effect on the path of the
object
6.
If the player manages to take the object to a tile with
a star then bonus score will be added
7.
If the object reaches the finish tile then the level is
completed and the score is given according to the
time taken and the number of tiles used
8.
Each tile will have a unique score to it
9.
Once the level is completed the next level will be
loaded
10.
Some levels will also have additional objectives like
1.
Speed traps
2.
Time limits
3.
Limited tiles etc.
6.3. Algorithm: Versus A.I.
Here each level will have a starting tile and a finish tile
1.
Each level will also feature some predefined tiles
that must be used by the player and 3 stars which
will grant bonus score
2.
The objective of the player is to take the object
from start tile to end tile and try to the beat the
score of made by the game A.I.
3.
The player will be allowed to choose from a set
of tiles to form a path to take the object from start
to finish
4.
Once all tiles are selected the game will choose
their own tiles with the A.I algorithm and show
them to the player
5.
After seeing the A.I. arrangement the player will
be allowed to alter any one of its tiles
6.
Once the player is ready both the object will be
released from start tile and pass through the
various tiles
7.
Each tile will have its own effect on the path of
the object
8.
The two objects will not collide with each other
9.
If the player manages to take the object to a tile
with a star then bonus score will be added but the
star will only be awarded to the object that first
reaches the star
10.
If both objects reach the star at same time then
both will get bonus score
11.
If the object reaches the finish tile then the match
is completed and the score is given according to
the time taken and the number of tiles used
12.
Each tile will have a unique score to it
13.
The score of A.I and Player is compared and the
winner will be declared as the one with the most
score
4
14.
Some levels will also have other criteria for
comparison like
1.
Speed traps at certain tiles
2.
Time Taken to finish
3.
Slowest to finish wins etc.
6.4. Algorithm: A.I. Algorithm
1.
Start
2.
For each type of tile make relations to all the
other tiles and store a compatibility score for each
relation
3.
All tiles will also have an initiating score which
is how suitable the tile is to start the game
4.
All tiles will also have special physics constraints
that will reduce misplacement of compatible tiles
5.
Using the initiating score the first tile will be laid
among the available tiles.
6.
After the first tile is laid, compatibility scores
will be used to find the most suitable three tiles
for the next step
7.
A random tile among these three tiles will be
selected
8.
After all tiles are selected the game will perform
a virtual run
9.
If the virtual run is unsuccessful then go to step 5
10.
If the virtual run is successful then show the tile
arrangement to the player and game will run on
these tiles to set score to beat for the player
VI. Mathematical Model
7.1. Problem Definition
Let `G' be the Game System that
G = {R, C, A, S, AI, B, Sp, Pr,T, Sc, IP}
R- Rendering engine
R = {R
1
, R
2
, R
3
,...Rn}
C- Collision engine
R = {C
1
, C
2
, C
3
,...Cn}
A- Animation
A = {A
1
, A
2
, A
3
,...An}
Engine may vary according to the unity S/W
S- Scripting {S
1
, S
2
, S
n
}
Also specify the language used
CF Core functionality = {R, C, A, S, AI}
AI Artificial Intelligence
D- Display along with GUI
Tile generator {t
1
, t
2
, ... t
n
}
L- Levels {L
1
, L
2
, L
3
, ... L
n
}
B- Board of size {n*m}
Speed- Sp = {Sp
1
, Sp
2
,... Sp
n
} depends on tiles
P- Period/ time of the Game i.e. Length of the
level
T- No. of tiles {t
1
, t
2
,... t
n
}
Pr - Parameters
It includes
Difficulty
Length
Obstacles
type
Pr = {D, L, O, T}
D- {D
1
, D
2
,...D
n
} Level of difficulties
L- {L
1
, L
2
,... L
n
} Length of the level of tile
O- {O
1
, O
2
,... O
n
} Obstacles in a tile
Type = T- {T
1
, T
2
,... T
n
}
Sc Scores
Sc- {Sc
1
, Sc
2
,...Sc
n
}
e.g. 3 star bonus
IP- Input -> touch
i.e. movement of touch
Left, Right, Up, Down, Teleport, Bomb, Ice,
Lava
7.2. Generalized Domain
It is a part of Reinforcement learning
On every time step the agent/ user must select an action in
discrete 3D action space.
e.g.: dim 1 {-1, 0, 1}
Used for left or right movement
dim 2 {0, 1}
Jump/ Not jump
dim3 {0, 1}
Walking/ Running
5
This 3D action space can be considered as action space of
distinct actions.
User receives state information from tiles as a 21 x 16 array
of tiles shows a state space corresponding to the view of
game.
Tile: Tells info about object. Can travel through tile, if it
can walk on top of the tile, etc.
Action value function Q -> rewards
7.3. Markov decision processes (MDPs)
Defines 4 tuples
MDP = {S, A, T, R}
S -> State
A -> Action
T -> Transition function : Game engine will take care of its
working
R-> Rewards: Game engine will handle its working
State:
User/ agent precieves the current state
SES (possible with noise)
Levels are episodic-
i.e. agent dies/ completes a level so it includes agent actions
it may or reset to level that is next
Action: A
Action available to the agent/ user
T: Transition function
T: S X A -> S
i/p = takes S & A as i/p & returns the state of the
environment after action is performed
Goal of agent=Maximize its reward/ scores. Scores are
represented using
Scalar values defined for reward function. R= S->R
T & S are controlled by game engine
7.4. User learning
Actions to take in state the policy.
: S->A
: Modified by learner over time to improve performance
defined as expected total reward/ score
Q: S X A -> R ( learning using temporal difference
method). Optimal action from any state by executing the
action with highest action value function.
As the state grows it can be stored in Table format of Q and
. But it is impossible to store them in the table form is the
state space is continuous using State variable
S = {x1, x2,...xn}
sSo this issue is resolved using tile coding, AI,
discretization
7.5. Morphism
In game "G" there are many objects interacting with each
other through functions. Here we make use of Petri net.
Petri net is a 3-tuple (B,E,F)
Where,
B= non-null set of conditions
E= set of events
F C= (B*E) U(E*B) is the causal dependency relation that
satisfy the restrictions
{b E B | e F b} are non-null for all events e E E.
Fig 4. Flow relations:
Definition:
Let N1= (B1, E1, F1, M1) be the nets for i= 0, 1
Defines a morphism of nets for N0 to N1 to be a pair of
relations (, ) with <= E0 x E1, a partial function and
B<= B0 x B1 such that:
1)
N1=N1 and
2)
If b0b1 then (`bo x `b1) is a total function
`bo b 1
3)
If e0 ex then (`ex *ey) is a total function ei->e0
And (b0' xb1') is a total function b0' b1'
If further is total we say the morphism (,) is
synchronous. When and are total functions we say the
morphism (,) is a folding. When and are the inclusion
relations :E0<=E1 and :B0 C=B1 we says N0 is a subnet
of N1.
6
Subnet provide the simplest examples of morphism on nets.
They have a simple characterization and arise naturally by
restricting a net to a subnet of events.
Proposition:
Let N0=(B0, E0,F0,M0) and N1=(B1,E1,F1,M1) be nets.
Then N0 is a subnet of N1if B0C= B1, E0 C= E1, M0 =
M1 and
Proposition:
Let N=(B,E,F,M) be a net. Let E' C= E. Define the
restriction of N to E', written N|E', to be (B, E', F',M)
where F'=F ((B x E') U (E' x B)).
The restriction N|E' is a subnet of N.
7.6. Activities
Activity A1: Start game
First, the player will select between predefined levels or
versus A.I mode
Activity A2: Assembly of engine with core functionality
Assembly of tiles according to levels
Activity A3: Agent vs. AI
Player will now play against AI
Activity A4: Validation of levels
All the tiles will be organized properly and game is set to
start
Activity A5: Score calculation
Scores will be calculated
Activity A6: Highest scores
Managing high scores of user
VII. Conclusion
Android as a full, open and free mobile device platform,
with its powerful function and good user experience rapidly
developed into the most popular mobile operating system.
Here we give an introduction of Android and developing
games for android.
Today developing android apps and publishing them has
been easier than ever. There are millions of android devices
in the market and are increasing by the day
MORE DEVICES = MORE DEMAND FOR APPS
Android games are today an important part of most peoples
leisure lives and increasingly an important part of our
culture as a whole.
VIII. References
[1] Sergey Karakovskiy and Julian Togelius, "The
Mario AI Benchmark and Competitions" member,
ieee, ieee transactions on computational intelligence
and ai in games, vol. 4, no. 1, march 2012
[2] Justin R. Martinez, 1 Wenbin "The Design and
Implementation of an Android Game: Foxes and
Chicken" Luo2 Vol.3, Issue.2, March-April. 2013 pp-
1129-1134
[3] Philipp Svoboda, Wolfgang Karner, Markus Rupp
"Traffic Analysis and Modeling for World of
Warcraft" reviewed at the direction of IEEE
Communications Society subject matter experts for
publication in the ICC 2007
[4] Jianye Liu Jiankun Yu "Research on
Development of Android Applications" 2011 Fourth
International Conference on Intelligent Networks and
Intelligent Systems
Frequently Asked Questions: 3D Android Game Development with A.I
What is the main topic of this paper?
This paper focuses on the development of a 3D Android mobile phone game using the Unity 3D game engine, incorporating Artificial Intelligence (A.I.). The game, called "Sammy," involves creating a path for a character to reach home from school.
What technologies are used in this game development?
The key technologies used are:
- Unity 3D game engine
- Android mobile operating system
- MAYA and Adobe Photoshop (for model and background design)
What are the key features of the developed game?
The game includes the following features:
- Two operational modes: single player and Vs. A.I.
- Three tiers of difficulty: easy, intermediate, and hard
- Intuitive user interface
- Smooth animation system
- Bright, vibrant colors
What are the game modes available in "Sammy"?
The game features two main modes:
- Predefined levels (single player)
- Versus A.I.
How does the scoring system work in the game?
The score is based on:
- Time spent to complete the level
- Number of tiles used
- Collecting stars (bonus points)
What algorithms are used in the game's development?
Algorithms are used for two game modes (Predefined levels and VS. A.I) and for A.I behavior.
What is the mathematical model used in this research?
The game uses a mathematical model based on:
- Markov Decision Processes (MDPs)
- Concepts of reinforcement learning
- Petri nets for modelling object interactions
What is the target audience for the game?
The game is designed for casual gamers, emphasizing simplicity, ease of understanding, frequent rewards, and high re-playability.
What are some future possibilities for the game?
By adding new game play features and tiles as the level progresses. Also new challenges can be designed in new updates.
Ende der Leseprobe aus 6 Seiten
- nach oben
- Arbeit zitieren
- Vrushab Randive (Autor:in), Shubham Vernekar (Autor:in), Chandrasen Kumakar (Autor:in), Monika Shirbhate (Autor:in), 2015, 3D Android Game Development with A.I, München, GRIN Verlag, https://www.grin.com/document/294682
Blick ins Buch
