Center for Intelligent Systems, Controls, and Robotics

Research in Intelligent Systems, Controls, and Robotics is a cooperative research effort in the automated systems area across four departments (Mechanical, Chemical, Electrical and Civil) in the College of Engineering. The aim is to develop implementable solutions to problems in systems, controls and robotics for applications in industry and government.

Intelligent Systems

Intelligent systems seek to improve the behavior of a dynamic system by using either computer implemented human reasoning (e.g., expert systems or fuzzy logic) or emulations of natural processes (e.g., neural networks or evolutionary algorithms). Research efforts in this area focus on the application of intelligent systems to robotics and other controlled systems. The following problems are being addressed:

• The development of automated controller tuning methods.
• The selection of weights to satisfy multiple objectives in optimal control design.
• The design of controllers for semi-active control of civil structures.
• The development of algorithms that allow a mobile robot to perform complex maneuvers among densely spaced objects.
• The development of fault detection and isolation algorithms to determine malfunctions in robots and other controlled systems.
• The development of classification algorithms for image processing in robotics and biomedical applications.

Controls

While most physical systems are nonlinear, most of the research in the past two decades has been limited to the design of linear controllers. However, with the development of high performance systems both in the military environment as well as the manufacturing sector, more stringent operational specifications have emerged for which the performance of linear controllers is inadequate.

Theoretical issues that are being addressed in this research are:

• The development of a diffeomorphism based on a nominal description of the dynamic system that results in an input-output system that is "almost" linear.
• The development of gain scheduling controllers to account for nonlinearities and time-varying parameters.
• The development of robust controllers to account for modeling uncertainty and disturbances.
• The development of a "separation principle" that will allow the independent design of gain scheduled and robust controllers.
• Characterization of the uncertain perturbations for the design approach proposed.

Practical problems that are being addressed to implement these theoretical ideas are:

• The development of numerical algorithms that can efficiently compute the controller outputs in real-time.
• The development of analysis tools for analyzing the efficacy of the proposed tools.
• The development of test-beds with the “hardware in the loop” to test the controllers.
• Practical implementation of the controllers in physical systems.

Robotics

The development of robots requires interdisciplinary skills from several branches of engineering and science. This research focuses on the development of mobile robots that perform assigned tasks in both industrial as well as military environments. In particular, the following issues are being addressed: