Human adaptive haptic sensing
File(s)
Author(s)
Carboni, Gerolamo
Type
Thesis
Abstract
How do humans physically interact with the environment or with other humans? It
is well known that the nervous system can modify the body’s stiffness by selectively
cocontracting muscles to shape the mechanical interaction with the environment, but how
this influences haptic perception is not known. This thesis examines whether humans can
adapt muscles’ activation to influence their perception of the physical interaction with the
environment. This question is investigated by conducting behavioural experiments using
dedicated robotic interfaces to study sensorimotor interactions in the presence of haptic and visual perturbations. Hypotheses about the underlying mechanism are then tested
through mathematical modelling and simulations.
Chapter 1 reviews related frameworks and introduces the most relevant questions addressed in this work. Chapter 2 then shows that the central nervous system (CNS) can voluntarily adapt muscle cocontraction to increase haptic sensitivity. In an experiment, participants tracked a randomly moving target with visual noise while being physically guided by a virtual elastic band, where the band’s stiffness was controlled by their muscle coactivation. The results show that participants learned to increase cocontraction with visual noise and decrease it when the guidance is incongruent with the visual target. The adaptation law governing the regulation of the body’s stiffness by the CNS is then derived through computational modelling. This model is designed to maximise visuo-haptic information
while minimising metabolic cost, thus trading off sensory information with energy.
Further, it is shown in Chapter 3 that when the subjects are coupled via a tuneable
connection to a robotic guidance designed to hinder their tracking through perturbations
at the turning points (where participants physiologically increase cocontraction), they
adapted cocontraction to reduce the impact of perturbations on performance. These results
highlight the CNS ability to modify the muscle activation patterns to improve performance
with minimal effort.
Chapter 4 tests the robustness of human adaptive haptic sensing introduced in the previous chapters for human-human physical interaction. For example, in tango dancing physical contact provides haptic information of the partner’s action required to coordinate the
movements. During such physical interactions, should one keep the arms compliant so that the partner can correct the motion, or should one stiffen them to better keep along the planned movement? Using a tracking task in which a dyad is coupled via a rigid connection, subjects readily adapted the compliance of their limb depending on both the accuracy of the partner’s and their own movement. The same computational
model introduced in Chapter 2 could explain these results and predict the experimentally
observed cocontraction adaptation. This suggests that the minimisation of prediction error and energy is a general principle also holding in interpersonal interactions.
Altogether, these findings shed light on how humans can adapt haptic sensing by changing
body properties, and propose a novel framework to interpret visuo-haptic perception for interaction with the environment and other humans.
is well known that the nervous system can modify the body’s stiffness by selectively
cocontracting muscles to shape the mechanical interaction with the environment, but how
this influences haptic perception is not known. This thesis examines whether humans can
adapt muscles’ activation to influence their perception of the physical interaction with the
environment. This question is investigated by conducting behavioural experiments using
dedicated robotic interfaces to study sensorimotor interactions in the presence of haptic and visual perturbations. Hypotheses about the underlying mechanism are then tested
through mathematical modelling and simulations.
Chapter 1 reviews related frameworks and introduces the most relevant questions addressed in this work. Chapter 2 then shows that the central nervous system (CNS) can voluntarily adapt muscle cocontraction to increase haptic sensitivity. In an experiment, participants tracked a randomly moving target with visual noise while being physically guided by a virtual elastic band, where the band’s stiffness was controlled by their muscle coactivation. The results show that participants learned to increase cocontraction with visual noise and decrease it when the guidance is incongruent with the visual target. The adaptation law governing the regulation of the body’s stiffness by the CNS is then derived through computational modelling. This model is designed to maximise visuo-haptic information
while minimising metabolic cost, thus trading off sensory information with energy.
Further, it is shown in Chapter 3 that when the subjects are coupled via a tuneable
connection to a robotic guidance designed to hinder their tracking through perturbations
at the turning points (where participants physiologically increase cocontraction), they
adapted cocontraction to reduce the impact of perturbations on performance. These results
highlight the CNS ability to modify the muscle activation patterns to improve performance
with minimal effort.
Chapter 4 tests the robustness of human adaptive haptic sensing introduced in the previous chapters for human-human physical interaction. For example, in tango dancing physical contact provides haptic information of the partner’s action required to coordinate the
movements. During such physical interactions, should one keep the arms compliant so that the partner can correct the motion, or should one stiffen them to better keep along the planned movement? Using a tracking task in which a dyad is coupled via a rigid connection, subjects readily adapted the compliance of their limb depending on both the accuracy of the partner’s and their own movement. The same computational
model introduced in Chapter 2 could explain these results and predict the experimentally
observed cocontraction adaptation. This suggests that the minimisation of prediction error and energy is a general principle also holding in interpersonal interactions.
Altogether, these findings shed light on how humans can adapt haptic sensing by changing
body properties, and propose a novel framework to interpret visuo-haptic perception for interaction with the environment and other humans.
Version
Open Access
Date Issued
2021-10
Date Awarded
2022-03
Copyright Statement
Creative Commons Attribution NonCommercial NoDerivatives Licence
Advisor
Burdet, Etienne
Publisher Department
Bioengineering
Publisher Institution
Imperial College London
Qualification Level
Doctoral
Qualification Name
Doctor of Philosophy (PhD)