A biomimetic MEMS-based tactile sensor array with fingerprints integrated in a robotic fingertip for artificial roughness encoding

Abstract

This work shows the accomplishment of a full integration of a biomimetic 2x2 tactile array and related electronics in an artificial fingertip. The technological approach is based on merging 3D MEMS sensors and skin-like artificial materials that are moulded mimicking human epidermal ridges. Experimental results using a mechatronic tactile stimulator for indenting periodic gratings (spatial periodicity from 400 μm to 1900 μm) and sliding them at constant speeds (from 5 mm/s to 40 mm/s) under regulated normal contact forces (between 100 mN and 400 mN) show that the developed sensing technology is suitable for fine roughness encoding: a frequency shift of the principal spectral component arising from sensor outputs was observed coherently with the spatial periodicity of the used ridged stimuli and their sliding velocity. Such phenomenon is pointed out with fine gratings particularly when the stimulation is operated along the proximal-distal direction of the finger (i.e. with sliding motion of the ridges of the stimulus across the ridges of the packaging) showing a more marked frequency locked behavior if compared to the radialulnar stimulation (i.e. with sliding motion of the ridges of the grating along the ridges of the packaging).Manuscript received July 31, 2009. This study was funded by the NANOBIOTACT project (EU-FP6-NMP-033287).C. M. Oddo is with the ARTS Lab

A mechatronic platform for human touch studies

“…The system was conceived to enable parametric, precise, repeatable and smooth stimulus presentation with standardized conditions, no vibrations and simple programming for indenting and sliding the surfaces on the fingerpad. Moreover, the chosen FPGA solution guaranteed to the platform adequate throughput together with design upgradeability [8,33], which was not typical of tactile stimulation systems previously reported in the literature [22,14,34]. Future works will present results of ongoing studies on the measurement of peripheral neural firing and brain responses by means of microneurography and EEG, and of (even combined) psychophysical experiments on human touch.…”

“…Future works will present results of ongoing studies on the measurement of peripheral neural firing and brain responses by means of microneurography and EEG, and of (even combined) psychophysical experiments on human touch. In parallel, the same system is enabling studies on artificial tactile sensors in robotics [8,33] and future investigations will focus on artificial touch evaluation of a variety of tactile surfaces, such as textiles.…”

“…1c) were applied with different indentation force levels to the index finger of a human subject to show sample protocols allowed by the open platform. To point out the flexibility in the upgrading of the chosen architecture, thanks to the used hardware-programmable logics, platform variables were transmitted to a PC via Ethernet at 5 kHz (significantly higher rates were possible) by a second parallel processor instantiated onboard the FPGA, similarly to [8]. Dynamic performance indexes, calculated on repeated runs of the same buffer of commands, are defined and provided in Section 4.…”

“…Actually, the continuum time integrals in Eqs. (8) and (9) are computed by means of a discrete summation, since the system is discrete time, with 5 kHz packet transmission frequency via Ethernet.…”

“…The robotic system has been devised with an open design approach since it is simple to command via a graphical user interface, is upgradable thanks to the FPGA control electronics, and can be used to perform neurophysiological studies in humans with techniques such as microneurography and EEG [7] even in combination with psychophysical experimental paradigms. Also, it is suitable for tribological and artificial touch studies as well and it allows to implement a wide variety of protocols for active [8] and passive studies [9].…”