Biomechatronic design of wearable and operational robots for rehabilitation and assistive applications

The development of efficient robotic systems for rehabilitation and assistive purposes requires the synergistic deployment of advanced solutions involving multiple aspects, such as the design of the kinematic structure, of the actuation system and a detailed understanding of the biological basis underlying recovery from neurological injury. This thesis investigates the application of biomechatronic design methods in two complimentary applications, namely the design of wearable robotic orthoses for gait assistance and the design of an operational robotic device for neurorehabilitation of the upper limbs. In the context of wearable robotic orthoses for gait assistance, an analysis of recent literature allowed to formulate a research hypothesis, which states that the choice of a non-anthropomorphic kinematic solutions for wearable robots can provide improvements, both from ergonomics and dynamical standpoints. However, the process of kinematic synthesis of non-anthropomorphic wearable robots can be too complex to be solved uniquely by relying on conventional synthesis methods, due to the large number of open design parameters. In order to address this design problem, this thesis describes a novel methodology, which allows to systematically explore the wide set of solutions provided by non-anthropomorphic wearable robotic orthoses and includes novel tests, specifically devised to solve the problem of enumeration of kinematic structures applied to a specified set of human segments and joints. Preliminary results, emerging from the implementation of the described methodology for the design of a hip-knee robotic orthosis, are reported to validate the described methodology. In the context of the design of robotic devices for neurorehabilitation of the upper limbs, the capability of a novel actuation architecture to guarantee a transparent interaction during patient-in-charge mode was demonstrated, based on an inverse dynamical model a Macro/Mini manipulator and on the analysis of experiments performed on 14 healthy subjects. However, many details concerning the relations between move- ment therapy, neural plasticity and recovery of motor function after stroke are still largely unknown: this is a recognized cause of limited efficacy of movement therapy. However, knowledge on this topic can be crucial for the design of a new generation of robotic devices for neu- rorehabilitation. To this aim, a preliminary study describing a novel integration of kinematic measurement technology with functional Magnetic Resonance Imaging (fMRI) to assess the neural correlates of motor recovery during robot-mediated therapy in chronic stroke is reported. This preliminary study was also functional to validate an experimental setup involving the execution of reaching movements in fMRI environments, providing a set of specifications, which were used for the preliminary design of a novel actuated fMRI-compatible robotic device.