SCEM Seminar

Over the last nine years, there has been a large, multi-university research effort, called CASA, centered at the University of Massachusetts Amherst to develop a next-generation radar system for quickly detecting and tracking severe weather phenomena such as tornadoes.

One of the results has been the development of a distributed adaptive radar system, called DCAS, which has been deployed in a 4-node test configuration in a region of the state of Oklahoma that has one of the highest incidences of tornadoes in the United States.

The test system’s performance has more than matched the original goals of the project. The system architecture is highly distributed and scalable, which brings up a number of interesting agent issues involving coordination of radar scheduling, distributed meta-level control of radar configurations, and multi-step optimization.

In this talk, I will first overview the basic architecture of the system, and then discuss my groups research on how these agent coordination and scheduling issues were solved in the system. This is joint work with Bo An, Sherief Abdallah, Yoonheui Kim, Mike Krainin, Chongjie Zhang together with Anita Raja and Shanjun Cheng from University of North Carolina Charlotte.