search

davesnowdon / nao-emotional-framework

A framework to allow NAO applications to share an implementation of an emotional model and provide a standard means of modifying rebot behaviour in response to emotional state.
Apache License 2.0
10 stars 5 forks source link

Do we need to represent "stance" #19

Open davesnowdon opened 10 years ago

davesnowdon commented 10 years ago

The one thing I didn't raise was 'stance' which is often used with VA to convey if the robot is wishing to physically engage or disengage with a stimulus eg by Dr Breazeal with Kismet. In PAD it's similar to dominance.

Initially I plan to infer it from VA eg if the robot is scared he will lean or move back. I've not seen anything in the literature about where stance values might really come from. I've not explicitly looked though as there is plenty to go at.

svenwoldt commented 10 years ago

Well I do not know much about your project of course,however, according to Dr. Lim, which you refer to sometimes.

https://www.youtube.com/watch?v=blLWpdGtYRQ

It is paramount to have Nao at least move the head all the time to make him look less angry. Or in the video she plays where Nao is holding his hands across his chest and playing with his fingers really gave him an adorable posture which evoked emotional commitment in the audience as you can see on the link. Maybe you have already seen it though.

davesnowdon commented 9 years ago

Sven, I agree that we at least need to change head position. Mike's python prototype code changes the head position according to the current emotional state, but we don't yet keep the head moving.

We decided back in October to move from using VA to PAD and so stance will be handled by the dominance value from PAD.

I'll close this issue once we've completed the move from VA to PAD