Study of control strategies for lower-limb rehabilitation exoskeletons

Rehabilitation exoskeletons are becoming a reality, promising to expand techniques and therapeutical methods to help people affected by mild to severe motor dysfunctions due to incidents, traumas, surgery, or aging; and to reduce the economic and social impact associated with these disorders. However, the offered flexibility, which would allow patients to use exoskeletal companions in daily life, must come to terms with the conflicting requirements, such as power-vs-weight and functional hardware design-vs-ergonomics. Despite the encouraging results, many researchers still agree on the prematurity of the field and, therefore, in designing exoskeletons without a specific market, human, and motor dysfunction target. This Thesis analyzes some control strategies implemented in exoskeletons for lower-limb rehabilitation. The design of these systems requires ensuring ergonomics, system stability, and performance. Contrary to the bipedal locomotion robots, whose literature is referred to for postural balance control law, a robotic exoskeleton is extensively in contact with the human wearer; therefore, it is required to guarantee stability when in isolation and once coupled to the user. Servo controllers, or “classic” regulators, cannot answer this request, but interaction controllers can by allowing to regulate the dynamic behavior of the systems at the port of interaction. Robotic rehabilitation devices must be able to assist patients during rehabilitation without undermining the neuroplastic benefits of the therapy. For this reason, the control law focuses on providing actuator torques to move the robot’s mechanical structure and partially assists patients by providing only the complementary forces required to complete the movement exercise. Further research investigated displaying a target dynamic behavior at the port of interaction, or coupling element, with the human to avoid patients’ discomfort or being affected by undesired effects due to exoskeleton dynamics. The main problems concern the apparent inertia displayed by the machine, which, since positive lower bounded, results in users being forced to exert net muscle torques not only to move their limb’s passive dynamics but also to drag the robot’s inertia. Fortunately, several control strategies have been proposed to solve this problem, such as the emulated inertia compensation. An extension of these techniques makes it possible to impose the exoskeleton to manifest a transparent behavior, i.e., make the coupled robot imperceptible to the human while they perform some movements, and making the robot intervene only to assist in solving unwanted situations, such as falling risk, or deviation from the pre-generated trajectory. Using mixed control techniques, by suitably tuning the transparent parameter, the physiotherapist can decide the amount of control authority left to the patient over the whole coupled system, according to their progress in therapy.