© 1999 Robert A. Freitas Jr. All Rights Reserved.

Robert A. Freitas Jr., Nanomedicine, Volume I: Basic Capabilities, Landes Bioscience, Georgetown, TX, 1999

7.4.2.3 Acoustic Inmessaging

A wide variety of sounds may be generated voluntarily by the patient and be detected by in vivo nanorobots, such as coughing, grunting, humming, various other mechanical body noises (Section 4.9.1.4) and even vocalizations including direct voice commands spoken aloud by the user (Section 4.9.1.5); 1-100 bit/sec inmessaging rates appear feasible by these methods. Intracochlear nanodevices can also be used for high-capacity inmessaging (Section 7.4.6.3).

Soundless subvocalizations may also be detected using kinesthetic techniques (Section 7.4.2.2) to measure ~0.5 mm movements in tongue position and the muscles of the larynx that produce adjacent distinguishable tones in the vocal cords. The triangular waveform harmonics emitted by the larynx are modified during passage through sub and supraglottal cavities, the shape of which may be voluntarily altered by submillimeter muscular movements -- again, well within the range that nanorobots can detect (Section 4.9.2.1). Data sharing by multiple nanorobots using a communication network should permit reconstruction and estimation of the intended vocalization with sufficient reliability to transfer useful noncritical data or commands into the network. In 1998, radar microchips operating at ~2 GHz were already used for soundless speech recognition -- the system recognized and interpreted motions of the human glottis, lips, tongue, jaw, and velum in real time.¹⁶⁴⁸

The field of voice recognition is active, and no systematic review will be attempted here.^{1654-1658,1710} In 1998, IBM's speech recognition technology, including the Voice Type Dictation and Personal Dictation systems (IPDS), was well-known and commercially available,^1659,1660 as was Kurzweil's VoiceCommands and VoicePro³¹³¹ and NaturallySpeaking from Dragon Systems.³¹³²

Last updated on 19 February 2003