Robust Performance of Nonlinear Systems

Jorge Tierno, Electrical Engineering, Caltech

An important step in the design of control systems, is the evaluation of the proposed design through simulation. In order to take into account uncertainty and noise, the simulation has to be repeated for a large number of values of the parameters and noise signals. In this talk we will present a simulation based method for computing a lower bound for the performance of a nonlinear system subject to noise and uncertainty. The algorithm is similar in nature to the power algorithm for the lower bound of $\mu$. Uncertainties allowed in the system are time invariant and time varying parameters as well as norm bounded unmodeled dynamics. We will also discuss how these approach relates to and improves upon the traditional methods for analysis.