MINIMUM MEAN-SQUARE ERROR STOCHASTIC LINEAR CONTROL,
Abstract
Optimal control algorithms are developed for discrete and continuous time stochastic linear dynamical systems for an 'instantaneous' weighted mean-square error performance measure. The derivations are based on well-known results in matrix analysis and the theory of stochastic linear systems. Cases where the system disturbance and/or the measurement error processes are each time-correlated are considered, and the original optimal control algorithms are modified to handle such cases. The results are attractive for on-line control applications since they require no precomputation. Stability analysis remains as a problem for future study. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Apr 01, 1968
- Accession Number
- AD0671047
Entities
People
- J. S. Meditch
Organizations
- Boeing