Neuromorphic haptic glove and platform with gestural control for tactile sensory feedback in medical telepresence applications

Abstract

This paper presents a tactile telepresence system employed for the localization of stiff inclusions embedded in a soft matrix. The system delivers a neuromorphic spike-based haptic feedback, encoding object stiffness, to the human fingertip. For the evaluation of the developed system, in this study a customized silicon phantom was fabricated inserting 12 inclusions made of 4 different polymers (3 replicas for each material). Such inclusions, all of them having the same shape, were encapsulated in a softer silicon matrix in randomized positions. Two main blocks composed the experimental setup. The first sub-setup included an optical sensor for tracking human hand movements and a piezoelectric disk, inserted into a glove at the level of the index fingertip, to deliver tactile feedback. The second sub-setup was a 3-axis cartesian motorized sensing platform which explored the silicon phantom through a spherical indenter mechanically linked to a load cell. The movements of the platform were based on the acquired hand gestures of the user. The normal force exerted during the active sliding was converted into temporal patterns of spikes through a neuronal model, and delivered to the fingertip via the vibrotactile glove. Inclusions were detected through modulation in the aforementioned patterns generated during the experimental trials. Results suggest that the presented system allows the recognition of the stiffness variation between the encapsulated inclusions and the surrounding matrix. As expected, stiffer inclusions were more frequently discriminated than softer ones, with about 70% of stiffer inclusions being identified in the proposed task. Future works will address the investigation of a larger set of materials in order to evaluate a finer distribution of stiffness values.