Iterative weighted least-squares identification and weighted LQG control design (Q1905921)

An adaptive control scheme is proposed based on iterative least squares identification in closed loop followed by a linear-quadratic Gaussian control design. The procedure takes modelling error into account during controller design and uses the closed loop controller requirements for determining the identification criterion. This is achieved for an \(H_2\) control problem by appropriate weighting of the least squares identification and LQG design problems. A simulation example is presented.