Abstract
As part of this thesis, our goal is to contribute to a better understanding of non-intentional rhythmical adaptation through a multidisciplinary approach combining experimental studies and computational modelling. This dissertation can be divided into two distinct and complementary parts. The object of the first one is the study of unintentional coordination during Human-Machine interaction. The purpose of the second one is to develop a computational model combining rhythmical entrainment and muscular control able to account for obtained experimental results as well as to simulate a motor behavior consistent with dynamics that can be observed during unintentional coordination.In the first experimental phase, we first studied the coordination between patients suffering from schizophrenia and a robot in the context of physical activity. We have made the choice to use this pathology as a tool to test and validate our first rhythmical entrainment model but also in order to better understand the origin of unintentional coordination. In the case of schizophrenia, patients suffer from an attentional deficit, social impairments, impaired intentional interpersonal coordination and impaired intrapersonal coordination. However, studies have demonstrated that non-intentional coordination during interaction with healthy subjects was preserved. Studying this pathology is therefore a mean to gain information on the role and the interaction of the different processes involved in the emergence of unintentional coordination. Moreover, the rhythmical entrainment model used to control the robot has been tested and validated during conditions bringing an ecological Human-Robot interaction (physical activity)In addition, in the experimental phase, we examined, subsequently, attentional processes related to unintentional coordination and more specifically we have focused on its attentional cost in healthy subjects. Thus, we conducted an experiment in order to evaluate the impact of unintentional coordination induced by a visual stimulus in the context of high attentional load. We have been able to assess the effect of unintentional coordination on attentional resources through different conditions.In the second part of this thesis, taking into account the literature review related to the modelling of motor control and the literature on interpersonal coordination enriched through our two experiments, we have developed a rhythmical entrainment computational model. We were interested in how rhythmical entrainment observed from a behavioral point of view could be translated in motor command sent to our musculoskeletal system in order to achieve coordinated actions. To do so, we have developed in simulation, an articulated arm model operated by sets of muscles (agonist/antagonist) controlled in force.