Two data-driven control algorithms for a MIMO aerodynamic system with experimental validation

This paper suggests data-driven Model-Free Adaptive Control (MFAC) algorithms for a Multi Input-Multi Output (MIMO) twin rotor aerodynamic system (TRAS). The TRAS is controlled using a MIMO control system structure with azimuth and pitch position control loops. A set of experimental results on laboratory equipment is included to show the performance comparison of the MFAC algorithms implemented both as a single MIMO controller and as two separate Single Input-Single Output controllers (in the two control loops) running in parallel, one for each axis of TRAS motion. The MFAC algorithms are compared with a controller designed by Virtual Reference Feedback Tuning and discussions of the measured results are given.

I. Introduction

Model-Free Adaptive Control (MFAC) is a data-driven control technique proposed and applied in [1] and [2] as MFAC algorithms. MFAC is advocated since it only relies on input/output (I/O) measured data of the process. Other advantages of MFAC are the stability accomplishment by means of resetting conditions and the general algorithm formulation in a Multi Input-Multi Output (MIMO) framework. Some successful MFAC implementations are concerned with the processes described in [3]–[10].