A fully-connected, distributed mesh feedback architecture for photonic A/D conversion

Abstract

We report a new approach to photonic ADC using a distributed neural network oversampling techniques, and a smart pixel hardware implementation. In this approach, the input signal is first sampled at a rate higher than that required by the Nyquist criterion and then presented spatially as the input to a two-dimensional error diffusion neural network consisting of M/spl times/N neurons, each representing a pixel in the image space. The neural network processes the input oversampled analog image and produces an M/spl times/N pixel binary or halftoned output image. By design of the neural network, this halftoned output image is an optimum representation of the input analog signal. Upon convergence, the neural network minimizes an energy function representing the frequency-weighted squared error between the input analog image and the output halftoned image. Decimation and low-pass filtering techniques digitally sum and average the M/spl times/N pixel output binary image using high-speed digital electronic circuitry. By employing a two-dimensional smart pixel neural approach to oversampling ADC, each pixel constitutes a simple oversampling modulator thereby producing a distributed A/D architecture. Spectral noise shaping across the array diffuses quantization error thereby improving overall SNR performance. Each quantizer within the network is embedded in a fully-connected distributed mesh feedback loop which spectrally shapes the overall quantization noise thereby significantly reducing the effects of component mismatch typically associated with parallel or channelized A/D approaches.