Location Modeling and Machine Learning in Smart Environments

1
Location Modeling and Machine Learning in Smart
Environments

• Robert Whitaker
• Supervisor A/Prof Judy Kay
• A/Prof Bob Kummerfeld

2
Overview

• Problem
• Previous Work
• Possible Data Sources
• Tools Available
• Issues

3
Thesis Topic

• Explore ways of determining a persons current
  location and activity
• Explore ways of predicting a persons
  location/activity using Location Modeling and
  Machine Learning
• The results returned must be scrutable

4
Possible Situation

• Wheres Boris Scenario
• Wish to organize a meeting with another person
  where the time suits both parties

5
Possible Steps

• Contact the person you wish to meet
• Both people would look at their schedules and
  negotiate a time
• Both parties agree on the time they are to meet

6
Possible Problems

• One of the persons schedule may be incomplete
• When you arrive at the meeting time the person is
  not there. Should you wait? Where is the person?
• What if you cant connect the person to organise
  the meeting

7
High Level View
8
Previous Work

• Active Badge Project
• Lancaster Guide Project
• Doppelganger
• Activity Compass Project

9
Active Badge Project

• First Indoor positioning system
• Users wear badges to emit their location
• Applied to teleporting
• Active Bat project extended the basic concepts
  developed

Source Nigel Davies and Hans-Werner Gellersen
Beyond Prototypes Challenges in Deploying
Ubiquitous Systems. IEEE Pervasive Computing,
Volume 1 (Jan-March 2002). 26-35.
10
Lancaster Guide Project

• A tourist guide for the city of Lancaster
• Used tablet PCs connected to a 802.11 network
• Limited by the infrastructure capabilities.

Source 1. Nigel Davies and Hans-Werner Gellersen
Beyond Prototypes Challenges in Deploying
Ubiquitous Systems. IEEE Pervasive Computing,
Volume 1 (Jan-March 2002). 26-35. 2. The Guide
Project, http//www.guide.lancs.ac.uk
11
Lancaster Guide Interface
Source The Guide Project, http//www.guide.lancs.
ac.uk
12
Doppelganger

• Generalized tool for gathering, processing and
  providing information about users
• Learning Techniques
• Beta Distribution
• Linear Prediction
• Markov Models
• DopMail

Source Orwant, J., Heterogeneous Learning in the
Doppelganger User Modeling System. in User
Modeling and User-Adapted Interaction, (1995),
107-130.
13
Doppelganger
Source Orwant, J., Heterogeneous Learning in the
Doppelganger User Modeling System. in User
Modeling and User-Adapted Interaction, (1995),
107-130.
14
Activity Compass Project

• Location Modeling to help disabled
• PDA device application developed to assist with
  location tracking
• Tracking movements and comparing them to a map
• Prediction algorithms used Relational Markov
  Models

Source Patterson, D.J., Etzioni, O. and Kautz,
H. The Activity Compass, University of
Washington, 2003.
15
Prototype of Activity Compass
Source Patterson, D.J., Etzioni, O. and Kautz,
H. The Activity Compass, University of
Washington, 2003.
16
Possible Data Sources

• Bluetooth Devices
• Machine Learning
• Windows Based
• Unix Based

17
Tools

• Personis
• Elvin Messaging
• Bspy
• Markov Modeling Toolkits
• Manual Logs for Evaluation Purposes

18
Personis

• User modeling software
• Accretion representation
• Consists of components which model aspects of the
  user
• Allows the user model to be scruntised

Source Kay, J., Kummerfeld, B. and Lauder, P.,
Managing private user models and shared personas.
in Workshop on User Modelling for Ubiquitous
Computing, (Pittsburgh, USA, 2003).
19
Example of User Model
Output from Personis Modeling the locations
where the user has been
20
Elvin Messaging

• Publish/Subscribe Messaging System
• Messages routed by content
• Application sending messages between sensors and
  modeling software

Source Mantara Software Elvin Administrator's
Guide, 2003.
21
Bspy

• Bluetooth positioning system
• Detects Bluetooth devices and logs them to a
  database
• Uses Elvin messages to send information from
  sensor to database

22
Example Data
23
Markov Modeling Toolkits

• Hidden Markov Modeling Package Python
• Matlab Hidden Markov Package
• Markov Chain Algorithm
• Cambridge Markov Modeling Toolkit

24
Manual Logs

• Records activity and location in 15 min blocks
• Provides some example data to develop the
  algorithms off
• Used for the evaluation of the learning algorithm

25
Code Sheet
26
Manual Log
27
Research Issues

• Representation of location and activity
• Creation of data sets
• Modeling Time

28
Questions