Tactile sensorimotor feedback strategies for object categorization in telepresence sensory augmentation applications

In this thesis we illustrate how vibrotactile stimulation, delivered via a haptic glove, can reliably represent objects properties like stiffness and texture. We describe how vibrotactile signals can provide sensorimotor feedback in telepresence applications, and their usefulness in tactile perceptual studies. The potentiality of vibrotactile stimulation is investigated within passive and active tactile tasks. In our first study, involving a passive tactile task, vibrotactile stimulation delivers information about object stiffness to a remote user, who is not performing any activity. We then perform two other studies, involving instead active tactile tasks. In the first one, vibrotactile information represents touch events or the forces exerted by a robotic arm while performing an industrial remote-control activity. In the second one, vibrotactile stimulation gives a tactile illusion of object textures, and is employed for the investigation of perceptual mechanisms of the human touch sense. We also report a detailed description of the design and implementation of the haptic device used to deliver vibrotactile stimulation, and we demonstrate its effectiveness in delivering reliable tactile information. Within the studies, haptic stimulation was performed following two main approaches. The first, performed for the telepresence studies, involved the implementation of neuromorphic models for the generation of spike trains, which activity was proportional to the sensed forces. The second approach involved the stimulation of the hand via sinusoidal waveforms. Results from our experiments showed how the implementation of neuromorphic spiking models results in a reliable representation of tactile information, and how vibrotactile stimulation can be employed for the investigation of the perceptual processes behind tactile perception. More in general, we demonstrated that vibrotactile stimulation can be largely employed for the representation of objects surface properties when delivered on the hands, thanks to the numerosity of parameters that can be tuned. As a conclusion, we can state that vibrotactile stimulation can reliably deliver information about objects properties when used in active and passive tactile tasks, with particular focus on telepresence applications. The acquired knowledge can be exploited for the development of haptic devices for sensory augmentation and substitution purposes.