A sensitivity approach to periodic control allocation in nonlinear systems

The objective of this paper is to propose a dynamic control allocation architecture for nonlinear systems that allows for the presence of periodic exogenous signals. The design is achieved by combining a gradient-based optimization strategy with the use of sensitivity dynamics for nonlinear systems that depend on parameters. The resulting strategy permits extending control optimization beyond the more standard point-wise (instantaneous) allocation schemes, as illustrated also by means of an academic example, while preserving the appealing output invisibility feature typical of the linear context.