• compensatory motions
• generalized control
• rehabilitation robotics

Rehabilitation devices, such as powered wheelchairs, assistive robotic arms,
and limb prostheses, offer significant assistance to individuals with residual motion
capabilities due to conditions such as stroke, disease, amyotrophic lateral
sclerosis (ALS), amputations, and spinal cord or brain injuries. The interfaces used
to control these devices (e.g., joysticks, head arrays, sip-and-puff systems, and electromyography
sensors) limit their effectiveness in controlling robotic devices with a
high number of degrees of freedom (DoFs) as the amount of human signals decreases.
Previous studies have explored the use of compensatory motions performed by prosthetic
users to control upper-limb prostheses with up to 2-DoFs. Building upon this
concept, I have introduced a novel methodology that is generalizable to multiple robotic
systems with a high and variable number of DoFs. Within this framework, it becomes
feasible to control several assistive robotic devices based on the residual movements
unique to each individual with reduced mobility. I employed inertia measurements
unit (IMU) sensors to read the motions performed by the user, intended to reach an
object, and used these signals to generate prosthetic joint commands according to the
user’s intentions. The robotic joints are then responsible for reaching the object. This
thesis presents the proposed control algorithm, describing its initial formulation and
development into a more general and valid law. I validated the algorithm through simulations
and experiments, starting from a prosthetic case and further expanding this
algorithm for the control of an avatar, such as the two-wheeled robot Alter-Ego. These
experiments with an avatar robot inspired me to conceptualize the robot as a sort of
“whole-body prosthesis”, wherein the assistive device is perceived as an artificial extension
of the user’s body. Consequently, I have further intensified my research on
exploring the applicability of this framework for assistive purposes, both in domestic
and clinical settings, with individuals experiencing reduced mobility, such as those
with spinal cord injuries or amyotrophic lateral sclerosis (ALS). Through these studies,
a generalized method is investigated that enables the control of multi-DoFs robotic devices
for assistance and rehabilitation purposes. This opens the door to a generalization
of the human-robotic device dynamic system, which can be applied in fields beyond
rehabilitation.