ECS-9702717 Walsh The objective of the research project is to create both simulation and design tools for the control of nonlinear systems. The educational objective is complementary, as it seeks to integrate the use of such tools, in the form of design and laboratory experiences, into the undergraduate and graduate curriculum where appropriate. The research will focus on two areas which lack unified procedures, The first area centers on controllable nonlinear systems with uncontrollable linearizations. Problems to be solved include path planning, tracking, and regulation. The second area is aimed at an emerging control system architecture in which the feedback loop is passed through a network. Such systems contain both asynchronous discrete events as well as a continuous nonlinear plant, and are best described as hybrid. Tools required include accurate simulation environments and design methodologies for tracking and regulation. Space-based robots, wheeled vehicles, underwater vehicles, and underactuated satellites (satellites with a reduced number of thrusters) are examples of controllable nonlinear systems with uncontrollable linearizations. Established techniques fail to stabilize such systems due to their underlying nonholonomic velocity constraints. A systematic procedure for solving the, path planning, tracking, and regulation problems is presented. Such procedures depend on the identification of a system with a normal form, where the problems are solved generically. Solutions involve optimal control for path planning, time-varying linear control about trajectories for tracking, and periodic forcing for regulation. While the author's previous work solves many example systems like the wheeled vehicles and some space-based robots, many important examples, for example the underactuated satellite and robots reorienting objects by rolling them between their finger tips, do not fit into the existing framework. The project concentrates on expanding the framework so that systems s uch as those mentioned above may be included in the generic theory. The second part of the research proposal focuses on networked control, which has emerged as a cost-effective architecture for large multiple-input, multiple-output control systems. Networked control technology has been applied with great economic benefit to automated manufacturing plants, process control, automotive control, and building environmental control systems. The author is currently studying the networked control of a blast furnace control system and a novel three-DOF translational platform. These control systems are inherently hybrid, consequently basic design and simulation tools are lacking. The project's objective is to develop a simulation environment and design tools, experimentally verified, which address this emerging industrial need.
