Answers are grounded in this paper and cite their sources.
It is shown that for a wide class of saturating control systems with random inputs, the optimum controller is an instantaneous non-linear function of the input phase coordinates and the output phase coordinates. If the system is a relay control system, the optimum controller is represented by a switching surface in the corresponding phase space. In special cases the optimum controller can be simplified by the use of error phase coordinates. These results are applicable when the random input is a generalized Markov process, a Gaussian process, or a Gaussian signal plus a Gaussian noise. The synthesis of the optimum switching surface by a self-optimizing technique is discussed.
No takes yet. Share an insight, caveat, or question.
A. T. FULLER (1960) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: