Professor P. S. Krishnaprasad
University of Maryland
There is growing interest in the technology and mathematics of
distributed control. The advent of smart nodes combining sensors,
actuators, local computation, and (wireless) communication poses
an interesting mathematical problem on approximate diagonalization
of plant matrices. Our answer (in joint work with George Kantor) is
based on gradient flows on matrix groups.
The availability of arrays of closely spaced MEMS actuators poses
interesting problems in the study of gradient flows on function spaces.
We note examples in adaptive optics in the published work of Mikhail
Vorontsov and his collaborators. We also discuss prospects for an
approach
(in joint work with Eric Justh) based on pattern formation with
applications
to the problem of controlling fields.