Models of processing systems only approximate physical reality. A major challenge addressed by our research is the design of controllers that, in spite of the uncertainty present in the model, can deliver high performance and also guarantee stable behavior. Such controllers are called robust. The effort focuses on the synthesis of robust controllers using methodologies such as pole-placement, predictive-control, and frequency-domain techniques based on the novel concept of a Nyquist Robust Stability Margin proposed by our group.

Since many processes of interest are inherently nonlinear, controllers designed using linear systems theory may experience severe performance degradation. Appropriate control designs that address relevant nonlinear features are needed. We are investigating the use of new methods for the analysis and design of adaptive, predictive, and sliding-mode control strategies, and the integration of the resulting techniques within the framework of computer aided tools.

We are developing technologies that enable the operation of control laboratory facilities in real time by students in geographically remote locations. The user accesses the laboratory using the Internet and a standard web browser, and can introduce manipulations and observe the response of the physical system. Applications of these research results include educational programs, as well as technology that allows the operation of complex chemical plants by personnel located in other parts of the world.