GTNavi System Hyojoon Kim, Sang Min Shim, Kai Wang, Pingping He CS8803 AIA, Spring 2009.

GTNavi System

Hyojoon Kim, Sang Min Shim, Kai Wang, Pingping He

CS8803 AIA, Spring 2009

Introduction

Previous / Related Work

New Features

Technology

Demo

Challenges

Evaluation

Future Work

Acknowledgement

Q & A

Outline

Do you know all existing buildings in Georgia Tech?

Do you memorize addresses of all of them?

Do you know all the exact locations, and how to get there?

What about visitors and guests to the campus?

Introduction

Lost in Campus..

Navigation application within GT campus

Not driving directions, but walking directions

Mobility – developed in Android

Additional features

Special instructions at special nodes (intersection points)

Friend Finder

New shortest path algorithm, etc

That’s why we need..

GTNavi System

Introduction

Static start point to end point

Driving direction only – tried walking, incomplete

Mostly need to know exact address

• Google Maps Driving Direction Service

Previous and Related Work

Web service – no mobility

Simple line from start to end point - no direction explanation

Not a real-time navigator

• GT walking path direction service

Previous and Related Work

1. Mobility

Mobile, handheld navigation device

No connection needed – easy deployment, high usability

Install, and you’re ready to go!

2. Walking path

People walking around campus

Guest and visitors, as well as you!

New Features

3. Flexibility

KML data – easy to update and deploy

Other campuses (it they have their own KML data)

4. Real-time Direction Service and More!

Left or right turn instructions

Images pops up as you approach

New Features

5. Wiki Page

Wiki page – user shares data about path

contribute to the system!

6. Friend Finder

Get the geo. Location of friend by SMS

The location will be set as destination

Shows the direction path to get to your friend!

New Features

1. KML data and Geo Points

KML file is a way of expressing geological points

XML style

Technology

2. Drawing on Google map

Extends Overlay Class

Technology

3. Database System

Data storage and management by SQLite

Technology

4. Finding the shortest path

Dijkstra's algorithm

Floyd–Warshall algorithm 

Complexity consideration

Technology

5. Wiki page

Why do we need it?

How do we want it to be?

Technology

1. Memory shortage

Android has limited memory

Dealing with memory efficiently

2. Runtime Finding path

Dijkstra  O(|V|2+|E|)

Floyd  O(length of the path)

3. Google Android G1

No sufficient document 

Getting used to the development environment

Challenges

KML Geo. Point data

Android SDK, Android Emulator

Google Map API

MySQL-Lite database

PHP to construct wikipage 

DDMS to debug

Resource

0 100 200 300 400 500 600 700 800 900 -

1,000,000

2,000,000

3,000,000

4,000,000

5,000,000

6,000,000

7,000,000

DB file sizeArray size

0 100 200 300 400 500 600 700 800 900 -

50,000

100,000

150,000

200,000

250,000

300,000

350,000

400,000

450,000

#of rows

Evaluation

0 100 200 300 400 500 600 700 800 9000

20

40

60

80

100

120

DB update time

0 100 200 300 400 500 600 700 800 9000:00:00.000

0:02:52.800

0:05:45.600

0:08:38.400

0:11:31.200

0:14:24.000

0:17:16.800

0:20:09.600

0:23:02.400

0:25:55.200

Algorithm running time

Evaluation

1. Better wiki page

User picking intermediate nodes

2. Faster performance

especially when showing all paths and buildings

3. Scalability

Georgia Tech campus is relatively small

Future Work

Professor Ling Liu

Matt Weber

Myungcheol Doo (congrats!)

Acknowledgements

Q & A