State Observer Implementation via Microcomputers with Applications
The microcomputer implementation of a Luenberger-type state observer for dynamic systems is discussed. For the linear case two types of implementations are considered, (a) when a discrete-time model is used, and (b) when the block-pulse function approach is utilized. Since the observations are invariably contaminated with noise, the former approach must include filtering the outputs before applying to the observer. On the other hand, the latter approach was found suitable even without a filter for a small amount of noise. Extension to the more general case of nonlinear systems is also discussed. The Luenberger-type observer can still be used, but the observer gains may have to be adjusted in an adaptive manner. Some results of simulation as well as a brief desscription of the hardware and the software are included.
KeywordsSampling Interval Extended Kalman Filter State Observer Gain Matrix Observer Gain
Unable to display preview. Download preview PDF.
- Azim, S.A. and Sinha, N.K., “Comparison of different methods of nonlinear state estimation with applications to a satellite”, 1978, Proc. 16th Annual Allerton Conf. on Communication, Control and computing (Monticello, 111.), pp. 202–212.Google Scholar
- Kalman, R.E. and Bucy, R.S., “New results in filtering and prediction”, 1961, Trans. ASME J. Basic Eng., sec. D, pp. 95–108.Google Scholar
- Li, M.H., “Nonlinear observer for the orbital states of an earth satellite”, 1980, M. Eng. Trans., McMaster University.Google Scholar
- Luenberger, D., “Observing the status of a linear system”, 1966, IEEE Trans. Mil. Electronics, vol. MIL-8, pp. 74–80.Google Scholar
- Sannuti, P., “Analysis and Synthesis of dynamic systems via block-pulse functions”, 1977, Proc. I.E.E., vol. 124, pp. 569–571.Google Scholar