Personalized Pedestrian Navigation Assistant: User Profile Assessment in PCRE Framework

1
Personalized Pedestrian Navigation Assistant
User Profile Assessment in PC-RE Framework

• Xiangkui Yao
• Graduate Research Forum
• Feb 27,2007

2
Talk outline

• Personalization
• PC-RE framework
• Pedestrian Navigation Assistant
• Hypothesis about personalized pedestrian
  navigation assistants
• Individual differences in spatial abilities
• Delivery and evaluation

3
Personalization

• Adjust and modify software configurations
  (functionality, interface, information content)
  based on users personal characteristics
  (abilities, needs and preferences)
• Importance of personalization

4
Personalization in adaptive user interface

• Bayesian network (Horvtiz et al, 1998 Liu et al,
  2003)
• Use probabilistic model to infer users
  goal/preference
• E.g. Microsoft Office Assistant (Clippit) based
  on Lumiere Project.
• Failure of Clippit difficulty to infer user
  goals
• Mixed-initiative (Liu et al, 2003)
• User and software agent take turns to initiate in
  accomplishing tasks
• E.g. LookOut System for MS Outlook (Horvitz,
  1999)
• Assume efficient collaboration between users and
  agents
• Model-based (Liu et al, 2003)
• No learning involved
• Need accurate user model
• Programming-by-example (Liu et al, 2003)
• E.g. Macro function in text-editing
• Substantial user efforts
• Difficulty in generalization

5
Personalization in e-commerce

• Major personalization techniques in web sites (Wu
  et al.,2003)
• Cookies
• Profile-based personalization
• Personal tools
• Opportunistic links
• Recommender systems
• Most techniques use machine learning and
  rule-based systems
• Troublesome when data instances are sparse
• (Saxe, 2004)

6
Personalization in RE 1

• Incorporate personal information requirements
  into design of Information systems for e-learning
  (Sun and Ousmanou, 2006)

Articulation process Personal Information
equirements PIR O x R x A x D x Pk x Sm A
access (judger or perceiver) O orientation
(introvert or extrovert) D decision on action
(feeler or thinker) R responsiveness to
information (sensing or intuitive) Sm sensory
modality (visual, auditory, or tactile) Pk
previous knowledge of the topic (poor, enough,
and good)
7
Personalization in RE 2

• Pros
• User-centered RE approach
• Take individual differences into account, and
  focus RE at individual level (Sun and Ousmanou,
  2006)
• Using a series of assessments that determine
  users abilities and attributes for learning the
  subject and prior knowledge
• E.g., one section of assessment questions for
  each preference (O, A, R, D, Sm, and Pk), using
  linkert scale
• Based assessment on domain theories
• Cons
• No implementation mentioned in their work and
  lack of evaluation
• Need to generalize the method for other domains

8
What we learn about personalization 1

• Accurate inference of users abilities/goals/prefe
  rences are difficult, especially at the beginning
  of usage
• Inaccurate inference could be counterproductive
• Alternative Assess users profile at the
  requirements engineering stage
• Particularly for systems that need to account for
  individual differences of ABILITIES
• E.g., Assistive Technology, learning and
  education applications

9
What we learn about personalization 2

• we can assess personal profiles and individual
  user requirements
• Based on domain theories
• Using tests and questionnaires
• Needs for evaluation (errors in personalization
  or customization could be counterproductive)

10
PC-RE framework

• Requirements engineering approach focusing on
  individual user and context
• Personal and Contextual Requirements Engineering
  Framework (PC-RE) (Sutcliffe et al, 2005
  Sutcliffe et al, 2006)
• Framework for personal requirements analysis
• Accommodate individual and personal goals
• Also, effect of time and context on personal
  requirements

11
PC-RE framework 1
Business domain evolution, user skills, expert
users
Layers
Temporal change
Culture localisation, interaction
language, style FRs
General stakeholder requirements
1
Spatial change
A
Individual user skill ability
Physical context, communications FRs, social
context
User characteristics, requirements
2
B
Attain individual goals
Location, social context
Personal goals
3
C
A Requirements specification B User model
characteristics C Personal goals and
preferences FR-functional requirements
Personal requirements framework and change
dimensions (Sutcliffe et al, 2005 Sutcliffe et
al, 2006)
12
PC-RE framework 2
Relations between the requirement layers
framework and system architecture (Sutcliffe et
al, 2005 Sutcliffe et al, 2006)
13
PC-RE framework 3

• The second layer (User characteristics
  /requirements) is of vital importance
• Traditional approach to assess group requirements
  of stakeholders in the RE process
• But, we need to assess users individual
  characteristics and requirements

(Sutcliffe, Fickas, and Sohlberg, 2005
Sutcliffe, Fickas, and Sohlberg, 2006)
14
Personal assessment the first step of
personalization

• Individuals take exam/test to have their
  personal profile assessed
• Requirements of the initial delivery of the
  system are based on personal profiles
• Use computer-based tests for such assessment, and
  have programs infer prescriptions automatically
• Assessment of personal characteristics are often
  domain-specific

15
Personalization under PC-RE

• Possible applications
• Learning systems
• Personalized health care system
• Personalized e-mail systems
• E.g. think-and-link for the cogntive impaired
• Pedestrian Navigation System

16
Personalized pedestrian navigation assistant

• Proliferation of mobile/wearable pedestrian
  navigation assistant systems both in the
  commercial market and research (Beeharee and
  Steed, 2006)
• Example scenarios
• Different PERSONAL spatial abilities -- mental
  rotation, visual memory, self-location, etc.
• Different CONTEXTS downtown, campus, shopping
  malls, etc.
• If we take a clinical and user-centered
  approach, we regard pedestrian navigation aids in
  such scenario as assistive technology

17
Existing pedestrian navigation assistants lacks
personalization

• Most existing pedestrian navigation assistant
  systems either targeted at the general population
  or focused on special groups, such as those with
  visual impairment or elderly population (May et
  al., 2003)
• Personalization is lacking or rigid in most
  existing systems (Baus, Cheverst, and Kray, 2005)
• Preferences vs. abilities

18
Need for personalized assistive technology systems

• Abandonment rates of AT systems are high (ranging
  from 8 to 75) (Kintch DePaula, 2002 )
• Part of the reason
• lack of respecting users characteristics
• (Kintch DePaula, 2002 )
• Individual differences in navigation

19
Individual differences in spatial abilities

• Large individual differences existing in human
  spatial abilities and strategies in spatial
  behavior (Hegarty et al., 2006 Kitchin and
  Baldes, 2002 )
• Individual differences in using external aids in
  navigation and learning map (Kitchin and Blades,
  2002)
• Individual differences in spatial cognition
  leads to strategy differences in navigation (and
  other spatial tasks) (Lobben, 2004)
• We need to take such individual differences into
  account

20
Research hypothesis

• Need for personalized pedestrian navigation
  asisstive devices
• Personalized navigation aid catered towards
  individual spatial abilities help user perform
  navigation tasks more effectively and efficiently
  than one without personalization
• To personalize, we need to have system
  requirements right for each individual!

21
Theories of spatial abilities 1

• Definition of spatial abilities (Golledge and
  Stimson, 1997)
• Geography definition
• the ability to think geometrically
• the ability to image complex spatial relations
  such as three-dimensional molecular structures or
  complex helices.
• the ability to recognize spatial patterns of
  phenomena at a variety of different scales.
• the ability to perceive three-dimensional
  structures in two dimensions and the related
  ability to expand two-dimensional representations
  into three-dimensional structures.
• the ability to interpret macro spatial relations
  such as star patterns or world distributions of
  climates or vegetation and soils.
• the ability to give and comprehend directional
  and distance estimates as required in navigation
  and path integration activities used in
  wayfinding.
• the ability to understand network structures.
• the ability to perform transformations of space
  and time.
• the ability to uncover spatial associations
  within and between regions or cultures.
• the ability to image spatial arrangements from
  verbal reports or writing.
• the ability to image and organize spatial
  material hierarchically.
• the ability to orient oneself with respect to
  local, relational, or global frames of reference.
• the ability to perform rotation or other
  transformational tasks.
• the ability to recreate accurately a
  representation of scenes viewed from different
  perspectives or points of view.
• the ability to compose, overly, or decompose
  distributions, patterns, and arrangements of
  phenomena at different scales, densities, and
  dispersions.

22
Theories of spatial abilities 2

• Psychology definition of spatial abilities
  (Golledge and Stimson, 1997)
• Spatial visualization
• ability to mentally manipulate, rotate, twist, or
  invert two- or three-dimensional visual stimuli.
• Spatial orientation
• the ability to imagine how configurations of
  elements would appear from different
  perspectives.
• Spatial relations (not clearly defined, include
  many things)
• abilities that recognize spatial distribution and
  spatial patterns
• identifying shapes
• recalling distributed phenomena comprehending
  and using spatial hierarchies
• regionalizing
• comprehending distance decay and nearest-neighbor
  effects in distributions wayfinding in
  real-world environments
• landmark recognition

23
Theories of spatial abilities 3

• Spatial ability tasks identified to assess
  individual spatial-related abilities/attributes
  in navigation (Lobben, 2004)
• Interpreting symbol meaning
• Route planning
• Self-locating
• Mental rotation of text/image/geometry
• Visual memory tasks
• Path integration

24
Delivery

• Prototype using CogBag system developed in
  Go-Outside project in UO Wearable lab.

25
Evaluation 1

• Personal profile assessment using computer-based
  tests and questionnaires using Navigational Map
  Reading Ability Test (Lobben, unpublished)
• Dimensions of spatial abilities
• Visual memory
• Mental rotation
• Self-location

26
Evaluation 2
27
Evaluation 3

• Wizard of Oz style experiment
• We have done this before for cognitively imparied
  population (Sohlberg, et al, to appear)
• Similar navigation tasks with same amount of time

28
Evaluation 4

• Compare performance of navigation using system
  with personalization and that without
  personalization
• Compare effectiveness (mistakes/finished tasks)
  and efficiency (time spent on tasks)

29
Evaluation 5
30
Evaluation 6
31
Conclusion 1

• major problem of existing personalization
  approach user models
• Alternative addressing the problem at the RE
  stage
• Individual profile assessment is the first step
  towards personalization

32
Conclusion 2

• Importance of differences between preferences and
  abilities in personalization
• Pedestrian navigation aid a good testbed for the
  idea
• The idea could potentially be applied to other
  applications where application domain theories
  demonstrate existence of great individual
  differences

33
Thank you for your attention! Questions?
34
Bibliography (1)

• Baus, J., Cheverst, K., and Kay, C. 2005. A
  Survey of Map-based Mobile Guides. Map-based
  mobile services - Theories, Methods and
  Implementations Meng/Zipf (Hrsg.), Springer, S.
  197-213.
• Beeharee, A. K. and Steed, A. 2006. A natural
  wayfinding exploiting photos in pedestrian
  navigation systems. In Proceedings of the 8th
  Conference on Human-Computer interaction with
  Mobile Devices and Services (Helsinki, Finland,
  September 12 - 15, 2006). MobileHCI '06, vol.
  159. ACM Press, New York, NY, 81-88.
• Fickas, S. 2005. Clinical Requirements
  Engineering. Invited paper at the 27th
  International Conference on Software Engineering
  (Extending the Discipline track), St. Louis, May
  2005.
• Golledge, R., and Stimson, R. 1997. Acquiring
  Spatial Knowledge, in Spatial behavior a
  geographic perspective. New York Guilford Press,
  1997, pp155-187.
• Hegarty, M., Montello, D. R., Richardson, A. E.,
  Ishikawa, T. and Lovelace, K. 2006. Spatial
  Abilities at Different Scales Individual
  Differences in Aptitude-Test Performance and
  Spatial-Layout Learning. Intelligence, 34,
  pp151-176.
• Horvitz, E., Breese, J. Heckerman, D., Hovel,
  D., and Rommelse, K. 1998. The Lumière project
  Bayesian User Modeling for Inferring the Goals
  and Needs of Software Users. Proceedings of the
  Fourteenth Conference on Uncertainty in
  Artificial Intelligence, July 1998.
• Horvitz, E., 1999. Principles of mixed-initiative
  user interfaces. Proceedings of CHI 99, ACM
  SIGCHI Conference on Human Factors in Computing
  Systems, Pittsburgh, PA, May, 1999.
• Kintsch, A., and DePaula, R. 2002. A Framework
  for the Adoption of Assistive Technology, SWAAAC
  2002 Supporting Learning Through Assistive
  Technology, pp. E3 1-10.
• Kitchin, R., and Blades, M. 2002. Individual and
  Gender Differences in Cognitive Mapping, The
  Cognition of Geographic Space, 2002, pp99-110.
• Liu, J., Wong, C. K., and Hui, K. K. 2003. An
  Adaptive User Interface Based On Personalized
  Learning. IEEE Intelligent Systems 18, 2 (Mar.
  2003), 52-57.

35
Bibliography (1)

• Lobben, Amy K. 2004. Tasks, Strategies, and
  Cognitive Processes Associated With Navigational
  Map Reading A Review Perspective. The
  Professional Geographer, 56 (2), 270-281.
• Lobben, Amy. 2006. Navigational Map Reading
  Ability Test. Unpublished.
• May, A., Ross, T., Bayer, S., and Tarkiainen, M.
  2003. Pedestrian navigation aids information
  requirements and design implications. Personal
  Ubiquitous Comput. 7, 6 (Dec. 2003), 331-338.
• Puerta, A. R. 1998. Design of Adaptive User
  Interfaces for Electronic Patient Records. In
  Proc. CHI 98 Workshop User Interfaces for
  Computer-Based Patient Records, 1998
  www.diamondbullet.com/cpr/paper-puerta.html.
• Saxe, R. S. 2004. Website Personalization using
  Data Mining and Active Database Techniques.
  Computer Science Seminar, April 24, 2004.
  RENSSELAER AT HARTFORD.
• Sohlberg, M. M., Fickas, S., Hung, P., Fortier,
  A. A comparison of four prompt modes for route
  finding with community travelers with severe
  cognitive impairments. Brain Injury, (to appear).
• Sun, L., and Ousmanou, K. 2006. Articulation of
  information requirements for personalised
  knowledge construction, Requirements Engineering.
  Volume 11, Number 4, September, 2006.
• Sutcliffe, A., Fickas, S., Sohlberg, M. 2005.
  Personal and Contextual Requirements Engineering,
  13th IEEE International Conference on
  Requirements Engineering, Paris, September 2005.
• Sutcliffe, A., Fickas, S., Sohlberg, M. 2006.
  PC-RE a method for personal and contextual
  requirements engineering with some experience,
  Requirements Engineering, Mar 2006, Pages 1 - 17.
• Wu, D., Im, I., Tremaine, M., Instone, K., and
  Turoff, M. 2003. "A Framework for Classifying
  Personalization Scheme Used on e-Commerce
  Websites," hicss, p. 222b, 36th Annual Hawaii
  International Conference on System Sciences
  (HICSS'03) - Track 7, 2003.