Closed loop parameter identifiability and adaptive control of a linear stochastic system (Q804537)

The author shows the parameters of a linear stochastic control system are identifiable when the system operates in a certain way under closed loop, if the parameters are identifiable when a persistently exciting input is used. It is assumed the controller depends on the test value of the parameters, and the control signal is dithered. The estimation problem is then formulated as a problem of minimizing an appropriate asymptotic cost function. By modifying the gradient of the cost function the problem is related to one in which Ljung's recursive scheme can be applied.