Asynchronous Event-driven Encoder with Simultaneous Amplitude and Temporal Fine Structure Extraction for Cochlear Implants
Student thesis: Doctoral Thesis
Related Research Unit(s)
Conventional cochlear implants (CIs) using periodic sampling are power consuming and incapable of capturing the amplitude and phase of the input acoustic signal simultaneously. This thesis presents an asynchronous event-driven encoder chip for CIs capable of extracting the temporal fine structure. The chip architecture is based on asynchronous delta modulation (ADM) where the signal peak/trough crossing events are captured and digitized intrinsically, which has the advantages of significantly reduced power consumption, reduced circuit area, and the elimination of dedicated data compression circuitry. An 8-channel prototype chip was fabricated in 0.18 μm CMOS process, occupying an area 0.125 × 1.7 mm2 and has a power consumption of 36.2 μW from a 0.6V supply. Comparator offset statistics are compiled from 30 silicon dies, showing very acceptable process variation. A 16-channel stimulation encoding system was built by integrating two test chips, capable of processing the entire audible frequency range from 100 Hz to 10 kHz. Experimental characterization using the human voice is provided to corroborate excellent functionality in the application environment.