Discrete Adaptive Control Allocation

TL;DR

This paper introduces a discrete adaptive control allocator for over-actuated systems that handles actuator uncertainties without needing uncertainty estimation, ensuring fast convergence and stability.

Contribution

It presents a novel discrete adaptive control allocation method that does not require uncertainty estimation or persistent excitation, with a closed-loop reference model for rapid convergence.

Findings

Effective control allocation demonstrated on ADMIRE system.

Fast convergence without oscillations achieved.

Robustness to actuator uncertainties shown.

Abstract

The main purpose of a control allocator is to distribute a total control effort among redundant actuators. This paper proposes a discrete adaptive control allocator for over-actuated sampled-data systems in the presence of actuator uncertainty. The proposed method does not require uncertainty estimation or persistency of excitation. Furthermore, the presented algorithm employs a closed loop reference model, which provides fast convergence without introducing excessive oscillations. To generate the total control signal, an LQR controller with reference tracking is used to guarantee the outer loop asymptotic stability. The discretized version of the Aerodata Model in Research Environment (ADMIRE) is used as an over-actuated system, to demonstrate the efficacy of the proposed method.