For the past several years Dr. Friedland, in collaboration with Dr. Avraham Harnoy of the Department of the NJIT Department of Mechanical Engineering, has been investigating the modeling and compensation of friction in control systems.

Dr. Harnoy and his students have developed dynamics models of friction in lubricated journal bearings based on the underlying physics and have experimentally verified these models. Current research is aimed at developing models for different types of friction and surface contact using a similar methodology.

Dr. Friedland and his students have developed a simple but effective technique for compensating friction in control systems based on use of a nonlinear observer to estimate the friction coefficient of a model based on simple Coulomb friction. Although the dynamics of the model are not used in the observer, it has been experimentally verified that the observer can track the friction dynamics with reasonable accuracy. Nevertheless, further performance improvements are expected to result from including the friction dynamics in the observer and estimating additional parameters.

In addition to the compensation of unwanted friction, Dr. Friedland and his students have initiated investigation of the modeling of friction in mechanical systems using frictional traction.

A nonlinear observer for estimating one or more parameters of a dynamic system was developed by Dr. Friedland as an extension of the technique used for estimating the friction coefficient. He and his former student, Dr. David Haessig, are now investigating the extension of this technique to the estimation of not only the parameters, but also the state of a dynamic system.

In his textbook, Advanced Control System Design, Dr. Friedland suggested the possibility of achieving a suitable control system design by "extended linearization" of the dynamics of a nonlinear system. Independently, Dr. James Cloutier arrived at the same method which he called the state-dependent algebraic Riccati equation method. A number of investigators have shown the validity of the method in examples, although a rigorous proof of stability has eluded discovery.

Dr. Friedland has shown that the method is capable of producing very good performance in systems with "hard" nonlinearities, such as friction, backlash, and limit stops.

Current research is being pursued on implementation issues to avoid on-line solution of the algebraic Riccati equation.