The Invisible Person

1
Towards engagingfull-body interaction

P.Petta1, A.Staller1, R.Trappl1,2, S.Mantler3,
T.Psik3, Z.Szalavári3, M.Gervautz3 HCI99
1Austrian Research Institute for Artificial
Intelligence Beliefs/Emotions in
2Dept.Med.Cybernetics Artificial Intelligence,
Univ.of Vienna Adaptive Interface Design
3Inst.of Computer Graphics, Vienna Univ. of
Technology
2
Summary

• principled background
• appraisal theory emotions as (grounded) process
• threefold use of emotions
• for uptake of stimuli
• for selection of action
• for expressive behaviour
• implementation
• appraisal integrated into thetrionic agent
  control architecture model
• radically simple scenario

3
Principled background appraisal theory

• the Emotional asconcern satisfaction process
• Concernscuriosity, familiarity,
  well-functioning, exercising, control,
  attachment,
• Appraisalsmediate between occurrence of
  concern-relevant events andchanges of internal
  action tendencies
• Action tendenciesstates of readiness to achieve
  or maintain a given kind of relationship with the
  environmentstimulus-driven, not goal-oriented

4
Threefold use of emotions stimulus uptake

• stimulus uptake at different levels
• in the context of the current behaviour
• in the context of the current activity (schema)
• in broader context (planning,)

5
Threefold use of emotions action selection

• action tendencies
• approach (desire), avoidance (fear),
  being-with (confidence),attending
  (interest), rejecting (disgust),
  non-attending (indifference), agonistic
  (anger), interrupting (surprise), dominating
  (arrogance), submitting (humility),
  deactivation (sorrow), bound activation
  (effort),excitement, free activation
  (joy),inactivity (contentment), inhibition
  (anxiety)
• as separate classification system for behavioral
  repertory of an agent
• behavior classescan be derived in a principled
  way froman analysis of a given
  lifeworld(including artificial ones)

6
3-fold use of emotions expressive behaviour

• emotion categorization by action tendency change
• vs.
• emotion categorization by nature of emotional
  object
• Highly dependent upon which objects
  aredistinguished and considered important
• Well-known example OCC model

7
3-fold use of emotions expressive behaviour

• ...emotions defined primarily by their object
  have no characteristic facial expression and ...
  their presence cannot be recognized by means of
  expressive behavior alone (N.Frijda)
• action expression problem as particular problem
  in architectures based solely on the second
  principle(e.g., the OCC Model)
• Action tendencies are tied tocharacteristic
  expressive behavior
• Behavior classes share specific expressive
  characteristics
• Emotions as changes in action tendency

8
Implementationlessons from control architecture
design

• Advances in particular with respect to layered
  designs
• Change in appreciation of the role of the
  middle tier of the established trionic
  model,the scheduler (middle-out model of
  control)
• scheduler receives important information
  fromlower level behaviors (cognizant failure)
  spawns long-term planning activities(cf. e.g.
  E.Gats ATLANTIS)
• scheduler as natural location for implementation
  of(a substantial part of) expressive action
  generation support

9
Implementation trionic agent architecture
Plan library (static)

• Communications
• vision data,animation status
• parameterizedanimation selection,emotion system
  state

Emotion subsystem action tendency dynamics

• Scheduler
• event appraisal,
• plan activation,
• RAP selection
• RAP execution monitoring

• Reactive action packages
• start/context/ending conditions
• idle/cleanup procedures
• cognizant failures

10
Implementation scenario (ALIVE, MIT)

• magic video mirrorwith back-projection
• video image mixedwith computergeneratedoverlay

• unencumbered interaction between human visitor
  and virtual character based on position,
  postures, and gestures
• main focus on virtual presence

11
Implementation constraints

• interface scenario issues
• avoid specialized scenarios,suggest
  straightforward, simple interaction
• balance sensorybehavioural capabilities ofthe
  synthetic actor andcapabilities elicited of
  human users

12
Implementation expressive behavior

• Animation is assembled in real time from basic
  motion captured sequences that are blended
• Dynamic surface texture synthesis in real time,
  based on current action tendency values

13
Implementation system architecture
VISION
appraisal ofuser location,posturegesture
AGENT CONTROL
Identification ofuser location,posturegesture
action tendency update
action selection
textureparameters
motion sequencing
segmentation
texture generation
motion blending
video imageacquisition
real-time renderingand compositing
ANIMATION