Information Transmission with a Multi-finger Tactual Display
4
Hong Z. Tan
Nathaniel I. Durlach
William M. Rabinowitz
Charlotte M. Reed
Proceedings of International Sensory Aid Conference, May 28-31, Sint-Michielsgestel, The Netherlands
1996 ISAC
http://hongtan.www.media.mit.edu/people/hongtan/hongtan-pub/conf/12ISAC96.ps
Motivated by the highly successful Tadoma method of speech
communication, a multi-finger positional display (the Tactuator) was
developed to study perception via the kinesthetic and vibrotactile
aspects of the tactual sensory system of the hand. The display
consists of three independent single contact-point actuators
interfaced (individually) with the fingerpads of the thumb, the index
finger, and the middle finger. Arbitrary stimulus waveforms from
absolute detection threshold to about 50 dB SL throughout the
frequency range from near DC to above 300 Hz can be delivered to the
three digits simultaneously (e.g., 25-mm low-frequency motion with
superimposed high-frequency vibration). Absolute thresholds measured
with the stimulator are in general agreement with results from the
literature.

The information transmission capabilities with the Tactuator were
assessed through a series of absolute identification experiments. Test
signals were derived by varying frequency and amplitude parameters of
multicomponent waveforms.  Three frequency regions were defined based
on their relatively distinct perceptual attributes: smooth motion (up
to about 6 Hz), a rough or fluttering sensation (about 10 to 70 Hz),
and smooth vibration (above about 150 Hz). Multi-component stimuli
were formed by summing sinusoids from each of these three regions,
with the intent that frequency and amplitude variations within each
region could be identified independently. Stimulation was applied to
either one of three digits (thumb, index, or middle) or to all three
digits simultaneously. For stimulus durations of 500 and 250 ms,
information transfer (IT) was 6.6 bits (corresponding to perfect
identification of 97 stimuli); at 125 ms, IT was 6.0 bits. Estimates
of potential IT rates were obtained by sequencing three random stimuli
and (a) having the subject identify only the middle stimulus and (b)
extrapolating this IT to that for continuous streams. Stimulus
durations of 125 to 500 ms and presentation rates of 1 to 7 items/sec
were tested. Estimated IT rate was about 12 bits/sec, and optimal
stimulus presentation rates were between 2 to 3 items/sec independent
of stimulus duration. This IT rate is roughly the same as that
achieved by Tadoma users in tactual speech communication.