Design, implementation and control of an electromechanical perturbation system for postural control evaluation

Orthostatic equilibrium conservation is one of the most critical features of the human body. Nevertheless, the biological control system behind it can be subject to reduced performance due to physical incidents, neurological pathologies, or aging. For this reason, recent researches on the biomedical field have focused on a proper method able to evaluate it in a univocal and replicable way for medical checkup reasons. Born as a collaboration between Politecnico di Torino Mechanics and Aerospace (DIMEAS) department and the Università degli Studi di Torino Neuroscience department, the Perturbation Generation Analysis System (PGAS) project revolves around the study of an alternative, more flexible way to evaluate human postural control. The system entails the use of a perturbative system, able to introduce a definite, unidirectional, and plannable source of disequilibrium on the subject’s body, and a rigid ground platform, to calculate the subject’s reaction to said distress. During the years, the perturbation system's various prototypes have focused on the biomedical correlation between the disequilibrium entity and its effects, but, as the PGAS research kept developing, the main problem has always involved the distress generation control. This is not only because manual or, later, pneumatic actuations introduce strong dynamic complications to the overall system, but also since human body reaction and general contact force dynamics are both highly complex, uncertain and nonlinear phenomena. This study aims to solve the control problem and refine the overall device design by implementing a linear electrical motor to the system, able to damp out or even delete many criticalities of the previous prototypes. The distress full characterization and design has revolved around planning a control algorithm that introduces stepped force profiles to body-machine contact interactions. This is because both force amplitude and impulse duration values were deemed critical to the study for a good biomedical characterization of the phenomenon. Considering that, as a basic approach to the problem, the system was approximated to a linear one, different straightforward closed-loop control methods were tested. Nevertheless, a serendipitous approach to the problem was deemed unable to adequately perform due to the electric motor's severe control input saturation. Therefore, the study has focused on a control algorithm able to consider this problem in advance, maintaining good overall performance, especially during the actuator's strike phase. As for the simulation and workbench testing, the used control algorithm has proved good first approach performances, but still requires robust control features further enhancements able to consider and damp out the many non-linearities and uncertainties involved in the phenomenon. Nevertheless, electromechanical actuation proved to be a far more efficient, straightforward and compact solution than the previous ones, and the prototype will serve as a good starting point system for future development.