Optimized Algorithm for Quasi-Quadrature Interferometric Signal Processing in Totally Implantable Hearing Aids

A new algorithm for processing of quasiquadrature interferometric signals in optical sensors of incus vibration, based on a 3x3 single-mode fiber optical coupler and low current VCSEL, is presented. This sensor serves as a microphone in the totally implantable hearing aid device. The amplitude of incus vibration spans over a large dynamic range, from a few picometers to a few nanometers, corresponding to sound pressure levels of 30-120 dB, superimposed on slow and hundreds micron large movement of the incus, which is induced by atmospheric pressure changes and the patient physiological activities. Instead of the usual interferometric phase demodulation based on arctangent function, which is processor time and power-consuming, we propose a simple, fast, but accurate technique. In this approach, the low-frequency components of the raw quasi-quadrature signals are used to determine the octant in which the photodetector signals are located and the audio-frequency component of one or another photodetector signal is forwarded to the output, after a simple processing. Two 16bit ADCs, a DC restoration circuit for each of the interferometric channels and one 16-bit DAC are employed. The sampling rate was 16 kS/s and an ultra-low power 16-bit fixed-point microcontroller performed the signal processing.