This book covers theoretical methods for the design of continuous time controllers for linear multivariable systems. It is intended for use by those wishing to build on a first course in control systems, either to expand their knowledge as practising engineers or as postgraduate students doing higher degrees in control engineering. It brings together state space and frequency domain design methods for linear multivariable systems. It compares their similarities and highlights the advantages or isadvantages of the various approaches for particular applications. The introductory first chapter is followed by five chapters dealing with state space design methods. Chapter two gives a detailed coverage of basic state space concepts including canonical forms, controllability and observability and minimal realisations. This is followed by a chapter on modal control which deals with both state and output feedback. A detailed coverage of quadratic optimal control is given in chapter four and this is logically followed by a chapter on observer design, which covers both full and reduced order observers, and shows how optimal control can be implemented when some states are not measurable.
The final state space chapter deals with decoupling contol, model following control and the robust seromechanism problem. Chapter seven is the first chapter dealing with frequency domain methods for multivariable systems and covers basic frequency domain relationships, system matrices and the Smith-McMillan form. The following two chapters cover the inverse Nyquist array and characteristic locus design methods and includes several examples to illustrtate applications of the methods. Chapter ten considers frequency domain design using factorisation techniques and contains an introduction to polynomial matrix methods. The final chapter discusses stochastic control problems begining with the classical Wiener filter, continuing with an overview of estimation theory and concluding with the Kalman filter.