Data-Driven Retrospective Cost Adaptive Control for Flight Control Applications

Abstract

Unlike fixed-gain robust control, which trades off performance with modeling uncertainty, direct adaptive control uses partial modeling information for online tuning. The present paper combines retrospective cost adaptive control (RCAC), a direct adaptive control technique for sampled-data systems, with online system identification based on recursive least squares (RLS) with variable-rate forgetting (VRF). The combination of RCAC and RLS-VRF constitutes data-driven RCAC (DDRCAC), where the online system identification is used to construct the target model, which defines the retrospective performance variable. This paper investigates the ability of RLS-VRF to provide the modeling information needed for the target model, especially non-minimum-phase (NMP) zeros. DDRCAC is applied to single-input, single-output and multiple-input, multiple-output numerical examples with unknown NMP zeros, as well as several flight control problems, namely, unknown transition from minimum phase to NMP lateral dynamics, flexible modes, flutter, and nonlinear planar missile dynamics.