Identifying Time-Varying pilot's responses: A regularized recursive Least-Squares algorithm

Mario Olivari, Joost Venrooij, Frank M. Nieuwenhuizen, Lorenzo Pollini, Heinrich Bulthoff

Research output: Chapter in Book/Report/Conference proceedingConference contribution

5 Citations (Scopus)


Methods for identifying pilot's responses commonly assume time-invariant dynamics. However, humans are likely to vary their responses during realistic control scenarios. In this work an identification method is developed for estimating time-varying responses to visual and force feedback during a compensatory tracking task. The method describes pilot's responses with finite impulse response filters and use a Regularized Recursive Least Squares (RegRLS) algorithm to simultaneously estimate filter coefficients. The method was validated in a Monte-Carlo simulation study with different levels of remnant noise. With low levels of remnant noise, estimates were accurate and tracked the time-varying behaviour of the simulated responses. On the other hand, estimates showed high variability in case of large remnant noise. However, parameters of the RegRLS could be further optimized to improve robustness to large remnant noise. Taken together, these findings suggest that the novel RegRLS algorithm could be used to estimate time-varying pilot's responses in real human-in-the-loop experiments.

Original languageEnglish
Title of host publicationAIAA Modeling and Simulation Technologies Conference
PublisherAmerican Institute of Aeronautics and Astronautics Inc, AIAA
ISBN (Print)9781624103872
Publication statusPublished - 2016
EventAIAA Modeling and Simulation Technologies Conference, 2016 - San Diego, United States
Duration: 2016 Jan 42016 Jan 8


OtherAIAA Modeling and Simulation Technologies Conference, 2016
CountryUnited States
CitySan Diego

ASJC Scopus subject areas

  • Aerospace Engineering
  • Modelling and Simulation

Fingerprint Dive into the research topics of 'Identifying Time-Varying pilot's responses: A regularized recursive Least-Squares algorithm'. Together they form a unique fingerprint.

Cite this