Affect in UAVs for Peace Keeping and Evacuation
Jill Greczek, DREU
Dr. Robin R. Murphy, advisor
Brittany Duncan, graduate student technical advisor
Our work is exploring how to use small unmanned aerial vehicles in close proximity to humans for peace keeping and evacuation. Imagine recent soccer riots. A small rotor-craft might be used to form a “swarm of angry bees” to keep people from rushing the field but then change into helpful “follow me” leaders for evacuation. We believe effective crowd control consists of the formation control (the movements) and the projected affect (the emotional response to the formation). Existing research shows that the affect projected by the UAVs is the key in persuasion, that the UAVs must be perceived as angry, trustworthy, etc. in order to be successful. However, almost nothing is known about how humans perceive affect in flying, non-anthropomorphic agents.
This project will
- research biological behaviors to determine a set of promising single and multi-agent formations and behaviors for crowd control (We have already identified behaviors where a single bird effectively protects its territory from predators on the ground),
- collaborate with subject matter experts in the performing arts at Texas A&M and Pixar to determine appropriate affective expressions such as sound pitch and frequency, acceleration and intensity of movements, and violations of personal space
- create a simulation of a single robot “blocking” and “herding” behaviors with affect.
- create a simulation of multi-robot “blocking” and “herding” formation behaviors with affect.
- Determine the appropriate simulator: MRDS, USARsim, Player-Stage, etc.
- Produce a conference-quality paper
The student needs to be familiar with artificial intelligence, affect, and UAVs.
The student would be responsible for the literature review, programming the simulations, and writing a report at the level of a conference paper. The student is expected to work closely with Brittany Duncan, a graduate student who is leading this research area, and the rest of the research team.
Summary: See a simulation of biologically inspired blocking and herding with 1 robot (with a simulation of N robots as a bonus)