Lecture 6 - Non Verbal expression for Robots

Question 1

Kinesics

Answer 1

Body movements, gestures, postures, facial expressions, eye gaze.

Question 2

Proxemics

Answer 2

Conscious or unconscious settings of distances between various objects, agents and ourself. Based on Hall’s proxemic zones.

Question 3

Hall’s proxemic zones

Answer 3

• Intimate (0-50 cm)
• Personal (0,5 - 1 m)
• Social (1-5 m)
• Pubic (5m or more)

Question 4

Haptics

Answer 4

Robot touch. Pushing buttons

Question 5

Chronemics

Answer 5

The tempo of human interaction and the pace at which we expect communication to occur.

Question 6

What is a non-verbal expression in robots

Answer 6

• Facial expressions
• Gestures
• Sounds
• Pose
• Body movements
• LEDs

Question 7

Arousal Vs Valence

Answer 7

Arousal is the degree of movement excitation an emotion shows.
Valence is how negative of positive an emotion is.

Question 8

Challenges of Body expression (7)

Answer 8

• Explicit body actions dedicated to communicating -emotions
• Interfere with functional behavior
• Difficult to use during task execution
• Emotion expression are not continuous over time.
• Between emotions, there is no state visible
• Affect can be expressed with task behaviors
• Mood indicates robots affect over time

Question 9

what is the model that varies behavior using parameters without interfering in the functions

Answer 9

Parameterized behavior model

Question 10

Naturalness

Answer 10

Behaviors generated by modulating APS are perceived as more natural than those generated modulating IPS

Question 11

Imitation Game

Answer 11

Study to model functional behavior through game gestures
Dyadic interaction in the lab with the idea of recognize robot mood.

Human players imitate feature sequences initiated by the robot after the robot finishes a sequence.

Question 12

Motivation for imitation game

Answer 12

Can people recognize robot mood expressed in behaviors during an interaction and can mood influence people’s affective state and performance during an interaction.

Question 13

storytelling motivation

Answer 13

To Express mood that is evolving over time in a multimodal communication context to improve storytelling experience.

Question 14

Storytelling Hypotheses

Answer 14

• Listeners can distinguish whether the robot’s body language expresses a mood that follows the story.

Question 15

StoryTelling findings

Answer 15

Affective body language is able to express mood evolving over time

• Consistent body language is important and makes the mood stronger
• Semantic channels take priority over body language.

Question 16

Lecture Experiment

Answer 16

Express mood overtime in an in the wild setting.

- Not a super useful experiment. no clear differences but the attention paid to lecture slides was higher.

Question 17

LED’s for emotions Motivation

Answer 17

• Minimal: Few parameters as possible
• Generic: Robot independent
• Parametric: Function of valence and arousal to cover full valence-arousal space

Question 18

Minimal set of parameters

Answer 18

• Motion features
  
  • speed
  • Amplitude
  • Repetition
• Body language Features
  
  • Head pose
  • Body pose

Question 19

What is affect in pose representation

Answer 19

A point in the valence-arousal space

Question 20

Amplitude

Answer 20

Valence oriented feature

Question 21

Speed

Answer 21

Arousal oriented feature

Question 22

8 affects for LEDs color

Answer 22

Excited, happy, content, tired, relaxed, sad, fear, angry

Question 23

LED’s parameters

Answer 23

• Color: Depends on valence and arousal.

- Blink Frequency is arousal oriented and also affects the “Affect”.

Question 24

Can people detect affects?

Answer 24

Depend:

• Motion only: excited, happy, sad, tired
• LED’s: Angry, relaxed
• Pose: Fear

Motion and body language features are strong indicators of arousal. Valence needs more features