Lecture 10 - Human Aware Navigation Flashcards
Name the three ways to apporach the human-aware navigation problem:
From a …… perspective:
- technological
- experimental
- technological
Rationale for human-aware navigation research
- robots operate more frequently in human environments
- for humans to feel comfortable, robots need to know how to navigate in an acceptable and comfortable way
concepts for human-aware navigation (name 4):
- comfort: absence of annoyance and stress
- naturalness: similarity between robot behavior and human behavior
- socialibility: adherence to cultural conventions
- safety: physical and perceived safety
What is the social equilibrium theory?
You can be close to someone when you make eye contact
What is proxemixs?
The study of spatial distances individuals maintain in social and interpersonal situations
What types or shapes of personal space do we create?
No vision in the back makes us feel uncomfortable
Elliptical space takes intended or current motion into account
Assymmetrical: taking the dominant hand into account
What are the three space types found in Kendon’s model?
O-space: shared interaction or focus of attention
P-space: group participant space
R-space: remainder
Experimental approach:
Main tasks for robots navigating in human environments (3):
- approach people
- avoid people
- do not occupy spaces used by people
What does the ultimate approach of a robot depends on (5, …)?
type of robot user group robot behaviour and capabilities activity of user personality of user overall context
Overall conclusions about distance:
comfort increases with distance
not circular: extended towards the back
no difference between left and right
results between two robots are quite comparable
Results about comfortable crossing strategies:
- crossing angle of -45* leads to a deviation to the left
- deviation is larger with a constant or increased speed change
- deviation s larger when the robot arrives earlier
- larger deviation leads to lower perceived comfort and higher perceived effort
Assumptions in particle filter simulation:
- known location and orientation of the person in the world
- known location and orientation of the robot in the world
- in simulation: known obstacles