Design and development of wearable devices and methods to study three aspects of motor redundancy: wrist posture, wrist impedance and supernumerary limbs

The Motor Redundancy of the Neuromuscular system is by far one of the most investigated aspects in Neurosciences and its effective and proper management, implemented by the Central Nervous System (CNS), is probably the key aspect that allows humans to safely, stably and gracefully move and physically interact with the environment. In this context, the present work aims at investigating three aspects related to the motor redundancy: i) managing the wrist redundancy during unconstrained kinematically-redundant tasks, ii) studying the wrist mechanical impedance during the physical interaction with the environment and iii) interfacing with redundant effectors, i.e. supernumerary robotic limbs (SRLs) by extending the concept of redundancy related to the human motor control, for augmenting human capabilities. Concerning the first scenario, we studied the control of the wrist and the management of the redundancy during the pointing with the wrist. It requires two degrees of freedom (DOFs) and previous studies have showed that the redundant DOF is bonded with the other twos on the basis of an empirical law, called Donders' Law. It is subject specific, path dependent and volatile (i.e. violations to this law have been reported in literature) and, geometrically, the Donders' Law represents a quadratic surface (also called Donders' Surface). In this work we further investigated this topic to study i) the stability over time of the Donders' Law and whether it remains during the motor adaptation to a visuomotor disturbance provided in the task space. Concerning the first scientific question, we carried out an experiment in which the enrolled subjects were asked to perform a pointing task in four different days and we found that the Donders' Law did not change over days within subjects, in terms of shape of the Donders' Surface. On the other hand, to study the effect of the adaptation to a visuomotor disturbance on the Donders' Law, the enrolled subjects were asked to perform a pointing task with and without the presence of a visuomotor disturbance provided in the task space. The main result that we obtained is the following: despite the subjects adapted to the disturbance, the specific implementation of the redundancy management (expressed in terms of shape of the Donders' Surface) did not change. Regarding the second scenario, in this thesis we presented the design and the validation of a portable wrist exoskeleton, conceived to be used in unstructured environments, to measure wrist stiffness/impedance and Parkinson's Disease related rigidity. The device consists in five revolute joints, of which four are passive and one is active; it allows providing mechanical perturbation to the wrist around Flexion-Extension. In a first experiment, we used the device to estimate the wrist stiffness both in healthy and in PD subjects. The device successfully allowed discriminating between i) healthy subjects and PD subjects and ii) PD subjects before and after the L-DOPA based treatment. In addition, we carried out a second experiment in which we validated the device in estimating the wrist impedance around Flexion- Extension, obtaining results congruent to the values reported in the literature. Lastly, concerning the third scenario, we presented a wearable device for providing vibrotactile feedback of a SRL. In a first experiment, we compared two feedback approaches considering unconstrained planar movements of the SRL: a kinematic feedback (i.e. the subjects received the end-effector position of the robot) and a dynamic feedback (i.e. the subjects received the torque of the two robot joints enabled). As result, we found the kinematic to be more reliable with respect the dynamic one, since it allowed subjects to correctly estimate the movements performed by robot. In a second experiment, we investigated the feasibility of a real-time estimation of the SRL's movement. To this aim, by providing the end-effector position of the robot, the enrolled subjects were able to correctly estimate the movement of the robot in real-time (while receiving the feedback).