Publisher's Synopsis
This book provides detailed consideration of high speed sampling issues in the control of linear continuous time systems using discrete linear controllers. It motivates the importance of controller synthesis from a signal processing view point, emphasizing the intersample performance and numerical considerations in controller implementation.;The features of the book include an emphasis on interaction between continuous time systems and the digital computers which implement the control algorithms based on a high sample rate measurement and a large number of numerical examples with routines available in MATLAB to allow further numerical investigations. The text confines most of the technical derivations to the problem sets, develops interpolating, decimating and block processing control algorithms based on signal processing techniques and develops explicit state space formulae for the discretization of analog controllers.;It presents new results of discretization based on closed loop performance; introduces properties of fixed and floating point arithmetic, and the consequences for the analysis of numerical errors in the implementation of digital control algorithms, describes a new dynamically scaled fixed point implementation; considers the importance of the state space structure for implementation of linear contollers in terms of minimizing the effects of errors due to finite coefficient and finite state wordlength.;New results on the structure of closed loop controllers are presented. The author also provides a complete self-contained derivation of the optimal Kalman predictor based on sampled measurements which are skewed with respect of the uniform pulse amplitude modulated control law and provides a complete self-contained derivation of optimal LQG regulator.
