White Noise Rejection in a Deterministic Setting

This paper provides a framework for analyzing white noise disturbances in linear systems. Rather than the usual stochastic approach, noise signals are described as elements in sets and the disturbance rejection properties of the system are considered in a worst case setting. The description is based on constraints in signal space, directly verifiable on experimental data. These constraints can be given a representation compatible with standard robust control, allowing the formulation of white noise rejection problems in the presence of other sources of uncertainty. It is also shown how the framework can capture as a special case the usual stochastic approach, with equivalent results.