Model Abstraction in Dynamical Systems: Application to Mobile Robot Control (Lecture Notes in Control and Information Sciences)

This book investigates abstraction of dynamical systems for the purpose of designing controllers. Abstraction is a means to reduce a system model's complexity while retaining the important behavior of that system. The motivating example throughout this text is the robotic car. Two topics are introduced in this text: control design using abstraction and propagation of uncertainty in abstracted systems. First, this book investigates the conditions for which controllers can be designed in abstracted systems and then transferred to the original dynamical system, taking advantage of design using the simpler model. This book also studies the relationship between the evolution of uncertain initial conditions in abstracted control systems. It is shown that a control system abstraction can capture the time evolution of the uncertainty in the original system by an appropriate choice of control input.