Patrick Olivier, Stephen Gilroy, Han Cao

1
Crossmodal attention public-private displays

• Patrick Olivier, Stephen Gilroy, Han Cao
• Daniel Jackson and Christian Kray
• Informatics Research Institute
• Newcastle University, UK

2
Talk summary

• public-private divide and ambient displays
• spatially context and location-based services
• crossmodal cognition and displays
• crossmodal ambient navigation (CROSSFLOW)
• extensions to display boards (CROSSBOARD)
• closing discussion

3
Public-private divide

• location-based services furnish information of
  different degrees of privacy
• handheld personal displays are the standard
  technological response to privacy issues
• personal displays
• place high demands on the user
• undermine our situational awareness
• need new display paradigms to address this

4
Ambient displays (design)

• subclass of peripheral displays
• low cognitive load
• match display environment (calm)
• sufficient information design
• consistent and intuitive mapping
• visibility of state
• aesthetic and pleasing design
• but they are not personalised

5
Information display spatial context

• standard configuration (location-based display)
• personal display
• positional sensing
• spatially multiplexed (specific location, all the
  time)
• alternative configuration (location-based
  display)
• global (spatially) display
• no positional sensing
• temporally multiplexed (specific time, all
  locations)
• user recieves a personal index to relevant
  information (time)

6
Crossmodal cognition

• multimodal cognition widely exploited
• capacities and effects involved in process of
  matching information received through multiple
  perceptual modalities (e.g. McGurk effect)
• humans spontaneously (and pre-attentively)
  integrate spatial cues across modalities
• GOAL utilise crossmodal signals to cue
  perception of temporally multiplexed information
  in a public display

7
Crossmodal ambient navigation
crossmodal cue
temporal multiplexing
crossmodal display
ambient display
8
(No Transcript)
9
Preliminary user study

• Goal
• compare CROSSFLOW a map on navigation task
• investigate impact of CROSSFLOW on a primary task
  
• explore the degree of ambience (cognitive load)
• Design
• dual-task
• primary task test a set of arithmetic questions
• secondary task find 5 of 15 targets using
  CROSSFLOW/map
• subjects 9 participants, 4 females and 5 males
• small experimental area 10 x 6.5 meters
• three phases for each subject

10
(No Transcript)
11
Observations
Dependent Variable Map CROSSFLOW Sample t p
Total time (secs) 133 80 M vs. C 3.5 lt0.01
Time per question (secs) 8.5 6.1 M vs. C 6.6 lt0.001
Questions correct () 84 98 M vs. C -2.3 0.054
Navigation errors 1.2 0.4 M vs. C - -
NASA TLX score 79 60 M vs. C 6.2 lt0.001

• the mean time 28 quicker
• the mean accuracy when using CROSSFLOW 17
  higher
• self-reports of cognitive load lower for
  CROSSFLOW

12
Discussion (on observations)

• crossmodal navigation led to improved performance
  on both the primary (arithmetic question
  answering) and secondary (navigation)
• directions appear vaguer in a small area with
  dense destinations than in a larger area with
  larger targets
• need evaluation of navigation in the wild
• questions
• are there other forms of temporal multiplexing to
  exploit?
• are there other situated displays to apply this
  to?

13
Crossmodal access to dense displays
14
Conclusions (on design space)

• modality design space
• modality selection configuration
• aggregation (e.g. visual continuity, dynamic vs
  static)
• public-private divide
• infrastructure (broadcast crossmodal cue
  schedules)
• action visibility / legibility (for users and
  bystanders)
• scalability
• temporal multiplexing inherently low resolution
• decompose the environment into regions (some
  tracking)
• evaluation and other configurations

15
Thanks for listening!