OPTIMUM DESIGN OF SAMPLED-DATA SYSTEMS WITH RANDOM PARAMETERS
Abstract
Three problem areas associated with the design of linear sampled-data systems are considered. The first arises from having the transition and distribution matrices of the system be random variables, i.e., the random parameter problem; the second from having multiplicative noise at the input to the system, this being a special case of the first problem area; and the third from being unable to measure the state vector of the system exactly. In each of these 3 areas, the preformance of the system is measured by using either a generalized sum-squared-error, a final-value, or a minimum-time criterion. The design procedures are based either upon minimizing the expected value of the performance index or upon minimiz'ng the performance index in the presence of worst-case variations within the system, e.g., minimizing the expected value of the sumsquared-error. In general, the results are in the form of feedback coefficients which relate the value of the optimum input to the value of the state vector of the system. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Apr 24, 1961
- Accession Number
- AD0255857
Entities
People
- T.l. Ii. Gunckel
Organizations
- Stanford University