Novel approaches to functional assessment and interaction control for robot-aided neurorehabilitation

It is commonly accepted that neuroplasticity is the basic mechanism underlying the recovery after brain injuries, and it has been widely shown that rehabilitation motor therapy can considerably influence recovery in patients with neurological diseases. However, the neurorehabilitation is labor intensive, often given through a one-on-one patient-therapist relation, and relies on manual interactions with a trained therapist several hours a day. In addition, patient evaluation is often done subjectively, with the therapist making hands-on or visual judgments about a patient's movement abilities. This makes it difficult to monitor and evaluate treatment effects. The introduction of appropriately designed machines could potentially enhance rehabilitation measurement by quantifying specific pathophysiological mechanisms, spontaneous recovery, functional ability, and therapy dosage more accurately than it is now possible. They could also help with therapy itself, replicating key components of current manual therapeutic techniques, or even applying new techniques, giving help in answering fundamental scientific questions and improving cost-efficiency of therapy. The main objective of this work is to provide innovative solutions to improve functional assessment in hemiplegic patients and enhance the outcome of neurorehabilitation by customizing the motor therapy on the basis residual motor capabilities of each single subject. To this purpose, this dissertation thesis presents, on the one hand, two algorithms to quantitatively analyze the clinical picture of patients with neuromuscular disorders; on the other hand, it tries to provide innovative control solutions that could enhance the performances of the existing machines used in robot-mediated motor therapy and tailor therapy interventions to each patient's needs and abilities.