Mario Olivari

Alumni of the Department Human Perception, Cognition and Action
Alumni of the Group Cybernetics Approach to Perception and Action
Alumni of the Group Motion Perception and Simulation

Main Focus

Research Scientist

I work in the and research groups.

My research interests include haptic guidance systems, pilot identification, motion cueing, and real-time model predictive control.

Background

I obtained my B.Sc. degree in computer engineering and my M.Sc. degree in automation engineering (both cum laude) from the University of Pisa. Between 2013 and 2016, I worked in a collaborative research project between University of Pisa and Max Planck Institute for Biological Cybernetics, resulting in a PhD for my work on identification of pilot control behavior.

Selected publications:
  • Identifying Time-Varying Pilot Responses: A Regularized Recursive Least-Squares Algorithm AIAA Modeling and Simulation Technologies Conference (2016)
  • Methods for Multiloop Identification of Visual and Neuromuscular Pilot Responses IEEE Transactions on Cybernetics (2015)
  • Pilot Adaptation to Different Classes of Haptic Aids in Tracking Tasks Journal of Guidance, Control, and Dynamics (2014)
Haptic Aid as Pilot Support System: a Human-Centered Design

Introduction

Haptic aids aim at helping pilots during a control task by providing force feedback on the control device. To assess whether the force feedback has a positive effect on the pilot performing the task, an insight is required into the pilot control behaviour. The dynamics of the neuromuscular system play an important role when trying to investigate the effect of haptic aids on pilot behaviour. Estimating the neuromuscular dynamics would provide quantitative insights into how pilots adapt their control behavior to the haptic aid.

Previous works have investigated methods to estimate the neuromuscular system during control tasks. However, a key limitation with much of estimation methods is that they make assumptions that may not hold for many practical applications [1, 3].

Goals

The goal of the project is to develop methods for estimating the neuromuscular system dynamics in compensatory tracking tasks with haptic aids, see Fig.1. The knowledge of pilot adaptation to haptic aids will be used to design the haptic aid based on pilot control behaviour (human-centered design).

Fig. 1 Tracking task with compensatory display and haptic aids.

Methods

We have shown that the commonly used method (CSD-based method) for estimating neuromuscular dynamics gives biased estimates in cases when a non-interference assumption is not fulfilled. We presented two different procedures, one based on ARX models (ARX-based method) and a multi-loop cross-spectral density method (CSD-ML method), which allow overcoming this limitation.

The two novel methods were validated with Monte-Carlo simulations and compared to the method commonly used in literature [1].  Furthermore, these methods were applied to experimental data obtained from closed-loop aircraft control tasks with pilot in-the-loop and with different haptic aids [2].

The CSD-based, ARX-based, and CSD-ML methods assume a time-invariant behaviour of neuromuscular response. However, the neuromuscular response is likely to be time-variant in realistic control scenarios, since pilots change their behaviour depending on environmental variables, fatigue, etc. Therefore, we have developed a method for estimating time-varying neuromuscular responses (RLS-based method) [3]. We are currently in the process of validating this method with experimental data.

Initial results

ARX-based and CSD-ML methods provided reliable estimates of time-invariant neuromuscular response, even when the commonly-used method failed [1]. Furthermore, results from pilot in-the-loop control tasks indicated that participants adapted their neuromuscular response Hadm to fully exploit different haptic aids, see Fig. 2 [2].

The RLS-based method gave reliable estimates of time-varying neuromuscular responses in a set of Monte-Carlo simulations [3].

Fig. 2. Neuromuscular responses for force-related tasks (force, relax, position) and for aircraft control tasks with different haptic aids (DHA, NoHA, IHA)

Initial conclusions

In this project, we showed that humans significantly adapt their neuromuscular response to the provided haptic aid. Future works will focus on designing the haptic aid based on the estimated neuromuscular responses.

References

1. Olivari M, Nieuwenhuizen FM, Venrooij J, Bülthoff HH, and Pollini L "Pilot Adaptation to Different Classes of Haptic Aids in Tracking Tasks", IEEE Transactions on Cybernetics, in press

2. Olivari M, Nieuwenhuizen FM, Bülthoff HH, and Pollini L (October 2014) "Pilot Adaptation to Different Classes of Haptic Aids in Tracking Tasks", Journal of Guidance, Control, and Dynamics, Vol. 37, No. 6 (2014), pp. 1741-1753

3. Olivari M, Nieuwenhuizen FM, Bülthoff HH, and Pollini L (October 2014) Identifying Time-varying Neuromuscular System with a Recursive Least-squares Algorithm: a Monte-Carlo Simulation Study, IEEE International Conference on Systems, Man, and Cybernetics (SMC 2014), IEEE, Piscataway, NJ, USA

Curriculum Vitae

Mario Olivari

Mario Olivari is a research scientist in the "" group and the group. His research interests include haptic guidance systems, identification of pilot control behavior, motion cueing, and real time model predictive control.

Current position

Since Jul '16 

Research Scientist at Max Planck Institute for Biological Cybernetics, Tübingen, Germany.

Education

Feb '13 - Jul '16

Ph.D. Candidate at Max Planck Institute for Biological Cybernetics, Tübingen, Germany (in collaboration with University of Pisa, Italy).

Thesis title: Measuring pilot control behavior in control tasks with haptic feedback.

Dec '08 - Dec '12

Master's Degree in Automation Engineering, University of Pisa, Italy.

Thesis title: Multi-loop pilot behaviour identification in response to simultaneous visual and haptic stimuli.

Final mark: cum laude
Sep '01 - Dec '07

Bachelor's Degree in Computer Engineering, University of Pisa, Italy.

Thesis title: Identifying linear systems with time delay by using Principal Component Analysis (PCA).

Final mark: cum laude

Experience

Nov '11 - Jun '12

Research Internship at Max Planck Institute for Biological Cybernetics, Department of Human Perception, Cognition and Action, Germany.

Project: Identifying pilot responses in control tasks with haptic aids

Dec '08 - Dec '09

Fellowship student at Italian National Research Council (CNR), Institute of Information Science and Technologies (ISTI), Italy.

Project: Investigating protocols for wireless communication

Dec '07 - Dec '08

Employee at University of Pisa, Italy.

Function: Computer and Network maintenance.

Honors and awards

Dec '16

Winner of the Best Dissertation Award 2016 from the Max-Planck-Institut für biologische Kybernetik and Förderverein für neurowissenschaftliche Forschung e.V.

This award is yearly awarded for the best Ph.D. dissertation of the Max Planck Institute for Biological Cybernetics.

Go to Editor View