w6 Measuring stress and social robots Flashcards
Why measuring stress in health care?
BPSD
Behavioural and Psychological Symptoms of Dementia
all behaviour that is associated with suffering or danger for the person with dementia or for persons in his/her environment.
Anxiety • Depression • Agitation • Aggression • Delusions/hallucinations • Apathy
people with dementia can’t communicate about their stress level, that’s why we need to technically measure it
Progressive lowered threshold model
explains the causes of BPSD as crossign the threshold of the “healthy level” of stress
people with dementia can’t communicate about their stress level, that’s why we need to technically measure it
how to measure stress with a wearable sensor?
movement, • temperature, • heart rate (variability), • skin conductance
some of the sensors are more or less sensitive to the movement interference, some of them the patients of nursing homes won’t wear
can we measure emotions with wearables?
Exact relationship between measurements and stress?
not that direct
AI for detecting stress with physiological measurements
False positives of stress detection
problems with wearables
Not recognizable (it’s a watch, but it isn’t)
Has to be tightened around the wrist (uncomfortable)
Power button is on the outside
Stakeholders do not like the design
Technical: PPG sensor fails during movement (cannot measure heart rate)
problems with wearables: what is preferred
Is the hippocampus primarily involved in emotional processing?
No
he’s more involved in memory and something about balance and also other things
Which term describes the continuation of an emotional response that is no longer appropriate?
Emotional perseveration
Rats with amygdala damage show no conditional freezing response. - T or F?
True, because they can’t experience fear without amygdala
emotional architecture of Kismet
Can people recognize Kismet’s facial expressions?
so seeing the moving robot (video) improves the recognition of emotions (that’s how we interact in real life, not through static pictures, but with living faces)
if people design the movements of an affective robot (like that breathing thing) manually it takes a lot of time and the repertoire of movements is modest, the patters repeat themselves. What to do?
Affective robot movement generation using generative adversarial neural networks (GANs)
Developing robot behaviors is difficult and time consuming
Automatic synthesis of affective robot behavior avoids repetitiveness
So they generated robot movement from human movement and the other way around
robots vs emotions
Theories on emotion serve as an inspiration for designing robot behavior
Robots (like humans) can express emotions using facial cues, physiological responses such as breathing, and body movements
Deep learning techniques can help to design affective robot behaviors
