EVALUATION COMMITTEE

The Faculty Council has appointed the following adjudication committee to evaluate the thesis and the associated lecture: 

• Prof. Stanisa Raspopovic, ETH Zürich
• Prof.  Christian Cipriani, Scuola Superiore Sant'Anna
• Assoc. Prof. Anderson de Souza Castelo Oliveira, Department of Materials and Production, Aalborg University, Denmark (Chairman).

Moderator:
Prof. Strahinja Dosen, Health Science and Technology, Aalborg University

ABSTRACT

A transradial amputation is a dramatic event that causes a substantial loss of motor and sensory functions. The state-of-the-art myoelectrically (EMG) controlled prostheses can be used to restore the lost motor function. However, despite the technological developments in mechatronics, many prosthesis users still abandon their devices due to several factors, from ergonomics and difficulties in control to the lack of somatosensory feedback from the prosthesis. The feedback can be restored using mechanical or electrical stimulation, but most of the studies in the literature focus on simple prosthetic devices with a single function (open/close). However, modern prostheses are advanced robotic systems that encompass multiple degrees of freedom, and therefore, several feedback variables need to be simultaneously communicated back to the user to effectively close the loop and convey the full state of the system. Electrotactile stimulation using surface electrodes is a non-invasive feedback strategy that can integrate many stimulation channels into a compact system. The aim of the present thesis was therefore to investigate how this approach can be used to convey multivariable feedback that is clear to perceive and easy to interpret.

To this aim, we investigated different feedback encoding schemes to represent the states of a multi-functional prosthesis and validated the feasibility of simultaneous electrotactile stimulation and EMG recording for stable closed-loop myoelectric control. The first study demonstrated that both spatially and intensity modulated feedback configurations intuitively conveyed discrete information regarding hand aperture and wrist rotation of a virtual myoelectric prosthesis controlled using EMG signals from the contra-lateral forearm to avoid interference. When placing stimulation and recording electrodes on the same arm, however, the stimulation pulses contaminate the recorded EMG-signal which disturbs the prosthetic control. Therefore, in the second study, a compact system for simultaneous recording and stimulation with integrated artefact blanking mechanism was presented  and the closed-loop control of a virtual prosthesis was further evaluated with the stimulation and recording electrodes placed ipsilaterally. The same setup was used in the last study to communicate the full state of a physical prosthesis (aperture, rotation, and grasp force) during functional prosthesis use. A novel mapping was proposed that generated tactile sensations which were anatomically congruent to prosthesis motions. The results from a functional task indicated that the novel feedback improved prosthesis control performance compared to the condition in which electrotactile stimulation was deactivated. In summary, the main contributions of the present thesis are the technical development of compact solutions for closed-loop prosthesis control, design of effective encoding schemes for multivariable feedback, and the assessment of human perception and benefits of such feedback. These are important steps towards clinical applications of modern prostheses enhanced with sensory feedback.