Pole assignment for stochastic systems with unknown coefficients (Q5926408)

The paper deals with a linear single-input stochastic system with unknown coefficients. In particular, the aim is to deal with methods of solution for the exact pole assigment problem under the controllability assumption, which is a necessary and sufficient condition for systems with known coefficients. To this end, the system noise is required to be mutually independent with zero mean and with bounded second moment. Two approaches are considered for constructing the methods of solution. The first method is a learning procedure and the second one belongs to adaptive class methods. Both methods require no coefficient estimation and are essentially based on stochastic approximation.