State-space self-tuning control with application to surface accuracy control of end milling

An algorithm is presented for a specific state of a locally linearized system to track any arbitrary external reference signal as closely as possible. The system may contain unknown parameters and the state need not to be measurable. The system (plant) is physically parameterized as a state-space model, and is assumed to be completely observable and controllable. The suggested algorithm can be viewed as a concatenation of the recursive prediction error method for joint state and parameter estimation of a state innovations model, and the state feedback control law with pole-zero placement in the transfer function from tracking error to reference signal. The practical usefulness of the algorithm is demonstrated by applying it to work-piece surface accuracy control in end milling.