On-line parameter estimation for indirect adaptive flight control: a practical evaluation of several techniques

摘要

To be able to maintain their performance, the Flight Control Systems in use on most transport aircraft should be endowed with resilient properties to deal with sensor/actuator faulty conditions or with any significant airframe changes. During the H2020 VISION project, several adaptive control designs have been developed by ONERA, and were flight tested to this end using the MuPAL-$lpha$ research aircraft of JAXA. One of these solutions involve a gain-scheduled controller designed off-line plus an estimation algorithm to track the time-varying parameters on-line, permitting the values of the main stability and control derivatives required to schedule the controller to be updated in real time. This paper focuses on the estimation part, and provides an experimental comparison of some results achieved in strictly similar conditions by using either frequency or time domain techniques. A new Hybrid Stabilized Recursive Least-Squares (HSRLS) is also proposed, as well as competing options regarding the aircraft modeling used by the estimator.