Usability in Pervasive Computing Environment Advance Usability October 18, 2004 Anuj A. Nanavati.

Usability in Pervasive Computing Environment

Advance UsabilityOctober 18, 2004

Anuj A. Nanavati

Agenda

Introduction Usability in pervasive computing Pervasive Computing (PVC) lab at IBM

Austin Lab overview Our work

Discussion issues

Introduction

"Things that think want to link", Nicholas Negraponte of MIT Media Labs

Nomadic/Amorphous/Ubiquitous computing User centered not machine or function centered Environment of connected computing devices Not personal computers but embedded or mobile

devices General purpose devices Communicating through interconnected network Remote access through internet

Introduction

Major playersPARC – origin of ubiquitous computing

research IBM – PVC LabMicrosoft – Easy Living MIT – Project OxygenNIST – Smart SpaceCMU – Project Aura

Usability in Pervasive Computing

Why usability testing? Understanding user interface User interaction with devices

Challenges in usability testing New environment for users No predefined user tasks Remotely moderated testing Wide range of devices – LCD on oven to PDA Tasks in multiple sessions and on multiple devices Testing tool should be independent of the device

being monitored

PVC Lab at IBM AustinLab overview

“Proof of concept” of networked areas Living room – TV, CD player, etc. Kitchen – oven, ScreenFridge, iron, etc. Garage – car Exercise room

Devices Devices common in above areas Control devices

Wireless touch control Web pad Connected through a service gateway

Physical Web site Every device has a URL

PVC lab overview

PVC lab overview

Technology Wireless technologyRadio Frequency tags – RFID Intelligent sensorsEmbedded devices

Our WorkGoals and Concepts – Dr. Bias

Goals Identify and fix usability problems in current UI Identify gaps in current usability testing methods

Concepts Remote migratory transactions

Start on one device complete on another Device independent UI

Start on PDA complete on ScreenFridge Task oriented UI compared to function guided

Supports the users in completion of their tasks Task categorization - taskonomy

Usability Testing

User interface for testing Engaged areas Engaged devices Recording options Remote views Other options - data logging, data analyses etc.

Sam’s Mock up: http://www.ischool.utexas.edu/~sburns/pervasive/#

Data collection requirements Observational data – Audio/video Biometric data

Usability TestingData Collection Requirements

Biometric data Eye TrackingGalvanic Skin ResponsesHeart RateBlood VolumeStress ParametersBlood Flow

Data Collection RequirementsBiometric data

Ubiquitous physiological monitoring Dr. Emil Jovanov from The University of Alabama at

Huntsville Prototype for Wireless Body Area Network

(BAN) of intelligent sensors (WISE) Personal medical monitoring Body sensors for brain, heart, movements etc. Sensors (WISE clients) controlled by a single

server (WISE server) connected to internet through a gateway

Data Collection RequirementsBiometric data

Taskonomy

Modeling user behavior in pervasive computing environment Categorize current user tasks into predefined categories to

facilitate UI design and usability testing methods: Assign future tasks to a category Assign devices to a category

Task categories User initiated vs. system initiated Single session vs. multiple sessions Single device vs. multiple devices Same vs. different device Single user vs. multiple users

Task scenarios Data downloading/uploading Heating food Setting up the temperature Many more….

Taskonomy

Structured representation of user tasks Task driven computing (2000), Zenyu Wang and

David Garlan, CMU Relieving Users from the Distractions of Ubiquity:

Task-Centered Architectural Framework (2002),  João Pedro Sousa, CMU

Task Description language (TDL)

Three step approach Based on Object Modeling

Taskonomy

Three step approach Identification

Devices Task categories Task scenarios

Connection Device device Device task category Device scenario

Modeling Scenario X for task category Y

Problems Defining scenarios Defining sub scenarios (actions)

D1

D3

D2

D5 D4

Links between devices show sub scenarios (actions) for a particular scenario of a task category

Taskonomy - Example

Scenario-1 (S1): Heat up the food kept in oven using your PDA through your work computer.

Task category-1 (T1): User initiated Ending on different device

Apply UI Design techniques (UID1) and Usability testing (UT1) methods to S1 and T1

Imagine in future similar scenario (S115) with same task category (T1) comes up. Apply UID1 and UT1 to S115

PDA

OVEN

WORK

S1 and T1

PDA

GAME

WORK

S115 and T1

Discussion Issues

Usability beyond traditional methodsUI beyond traditional UIModeling user behavior in pervasive

computing environmentConcerns

Privacy Security

References

IBM’s PVC LabCandace A. York, Senior software engineer, IBMhttp://www-106.ibm.com/developerworks/wireless/library/wi-pvc/

Ubiquitous and pervasive computing resources

http://www.iturls.com/English/TechHotspot/TH_77.asp Stress monitoring using a distributed wireless intelligent sensor system 2003, Emil Jovanov, IEEE Task driven computing 2000, Zenyu Wang and David Garlan, CMU Relieving Users from the Distractions of Ubiquity: Task-Centered

Architectural Framework 2002,  João Pedro Sousa, CMU

Thank You