Speaker: Dr. Azad M. Madni is Technical Director of the Systems Architecting and Engineering Program and a Professor in Astronautical Engineeering Department at the University of Southern California.
Abstract: Distributed autonomy pertains to a sociotechnical system in which geographically distributed human and software agents make independent observations, pool them to jointly determine the state of the environment, and then make collaborative decisions based on their shared goals and knowledge of environment state. This Talk presents ongoing research on adaptive cyber-physical-human system, an example of intelligent distributed autonomy. The perimeter security of a parked transport aircraft using multiple surveillance quadcopters is used as an illustrative example. The distributed system simulation comprises: a world server, which maintains the state of all world entities and runs continuous dynamic simulation of all quadcopters; a perimeter defense computer, which runs the dashboard controlling the quadcopters; and an opposing force computer, which runs the dashboard controlling the enemy unit. In this simulation, decision making is also distributed with sensor fusion agents (that determine environment state from sensor data) residing on the world server, and decision-making agents (that exploit knowledge of environment state information to make decisions) residing on the perimeter defense computer.
For more information, visit: https://sercuarc.org/event/serc-talks-whats-really-distributed-in-distributed-autonomy/