Reduced-Order Sensitivity Models of Linear Systems with Applications.
ILLINOIS UNIV URBANA COORDINATED SCIENCE LAB
Pagination or Media Count:
The study is directed towards developing reduced-order parameter and initial condition sensitivity models for linear time-invariant continuous-time, sampled-data and time-delay systems. By introducing the concept of companion relationship among sensitivity functions, it is established that the sensitivity system os a general cyclic continuous-time system has the same dimensions whether the system is represented in companion form or in phase-variable form. Computational aspects aspects for attaining higher accuracy and greater efficiency in sensitivity realization are discussed. For systems with real eigenvalues, it is shown that its sensitivity system can be realized efficiently using Jordan canonic form. Generation of input parameter sensitivity functions is studied and reduced-order models for two classes of input functions are developed. Author
- Theoretical Mathematics