Robust Control Theory and Applications
Abstract
This AASERT program supported research in robust simulation, hierarchical uncertainty representation, and novel methods for robustness analysis of uncertain systems. In the context of this program, robust simulation means simulating simultaneously sets of initial conditions and disturbance or noise signals. Thus sets of state space must be propagated by the dynamics of the model. Initial investigations have focused on piecewise linear discrete time systems, which map polyhedra to polyhedra at each time step. Linear programming can be used to refine the resulting bounds. This is important if the potentially exponential growth in set descriptions is to be overcome. Hierarchical uncertainty modeling is a new framework to include explicit representation of uncertainty in component modeling. The focus has been on LFTs and implicit (DAE) representations. A variety of examples including parasitics and non linearities illustrate the key ideas. Finally, this report describes new bounds on a spherical mu problem that allows for correlations between uncertainties in an LFT framework. Interestingly, this setting provides quite elegant bounds and simplified computation.
Document Details
- Document Type
- Technical Report
- Publication Date
- Feb 06, 1998
- Accession Number
- ADA337888
Entities
People
- John Doyle
Organizations
- California Institute of Technology