Barry L. Shoop

Optical Oversampled Analog-to-Digital Conversion

A new approach to optical analog-to-digital (A/D) conversion based on oversampling and interpolative coding techniques is described, and both interferometric and noninterferometric architectures based on this method are presented. This new approach combines the high resolution of classical oversampled A/D conversion with the high speed of optical processing technology to extend the resolution and conversion rates beyond that currently possible with other electronic or optical converters. The proposed optical converters are simple, consisting of multiple quantum well self-electro-optic effect devices, photodetectors, and common optical components that are capable of operating at sampling rates of up to 15 Gbits/s and that can provide scalable resolutions of 16 and 8 bits at a conversion rate of 117 and 938 MHz, respectively.

Noninterferometric optical subtraction using reflection-electroabsorption modulators.

A noninterferometric technique for optical subtraction is demonstrated that employs a multiple-quantum-well reflection-electroabsorption modulator and provides lower insertion loss, larger contrast ratio, and linearity over a larger dynamic range than similar techniques previously reported.

A first-order error diffusion modulator for optical oversampled A/D conversion

Abstract A first-order error diffusion coding modulator using multiple quantum well modulators for use in an optical oversampled analog-to-digital converter is demonstrated for the first time. The modulator was operated at optical sampling rates up to 1 kHz and demonstrated performance metrics within 2% of those predicted by theory.

High resolution optical A/D conversion using oversampling and interpolative coding

A novel approach to optical analog-to-digital (A/D) conversion based on oversampling and interpolative coding techniques is proposed. This approach incorporates the high resolution capabilities of classical oversampled A/D conversion with the high speed of optical processing technology to extend the resolution and conversion rates beyond those currently possible with other electronic or optical converters. The proposed optical converter uses optical threshold and arithmetic operators and fundamentally trades bandwidth for improving amplitude resolution. A non-interferometric realization using multiple quantum well electroabsorption modulators, photodetectors, and a novel approach for noninterferometric optical subtraction is described that can provide 8 bit resolution at a conversion rate of approximately 1 GHz.<<ETX>>

Analog-to-digital conversion using single-layer integrate-and-fire networks with inhibitory connections

We discuss a method for increasing the effective sampling rate of binary A/D converters using an architecture that is inspired by biological neural networks. As in biological systems, many relatively simple components can act in concert without a predetermined progression of states or even a timing signal (clock). The charge-fire cycles of individual A/D converters are coordinated using feedback in a manner that suppresses noise in the signal baseband of the power spectrum of output spikes. We have demonstrated that these networks self-organize and that by utilizing the emergent properties of such networks, it is possible to leverage many A/D converters to increase the overall network sampling rate. We present experimental and simulation results for networks of oversampling 1-bit A/D converters arranged in single-layer integrate-and-fire networks with inhibitory connections. In addition, we demonstrate information transmission and preservation through chains of cascaded single-layer networks.

Optimal error diffusion for digital halftoning using an optical neural network

A novel technique for digital image halftoning is proposed based on a symmetric error diffusion algorithm and an optical realization of a neural network. Using this approach, all pixel quantization decisions are computed in parallel and therefore the diffusion filter need not be causal. Visual artifacts resulting from the causality of the diffusion filter are reduced and therefore halftoned image quality is improved. Also, the inherent parallelism associated with optical processing can reduce the computational requirements while decreasing the total convergence time of the halftoning process.<<ETX>>

Second-order cascaded optical error diffusion modulators for oversampled analog-to-digital converters

Abstract In an oversampled A/D converter, oversampling, negative feedback, and linear filtering techniques are used to trade sampling speed for improved amplitude resolution. In order to achieve higher resolution with this type of converter, a higher-order noise shaping filter is required. However, as a result of the negative feedback architecture, higher-order filters can cause limit cycle oscillations and therefore instabilities in the converter. Cascaded oversampling modulators offer a means of achieving higher resolution without the stability problems associated with single-loop modulators. The feasibility of a two-stage cascaded optical oversampled modulator is investigated and found to provide second-order performance with relatively modest constraints on stage-to-stage matching tolerances.

An error diffusion neural network for digital image halftoning

A novel technique for digital image halftoning is proposed, based on a symmetric error diffusion algorithm and a new form of artificial neural network. Using an error diffusion neural network, all pixel quantization decisions are computed in parallel and therefore visual artifacts resulting from the causality of the diffusion filter in classical error diffusion techniques are reduced and the resulting halftoned image quality is improved.

Laser-power stabilization using a quantum-well modulator

A novel method for laser-power stabilization using a multiple quantum-well reflection electroabsorption modulator with a Fabry-Perot cavity is demonstrated. Stable operation of the modulator produces a linear relationship between control current and optical absorption which permits the realization of noninterferometric optical subtraction and the implementation of a negative feedback architecture. Using this stabilization technique and a modulator with a 70% reflectivity change, intensity noise as large as +or-75% of the incident level can be eliminated from a noisy laser beam.<<ETX>>

Precluding nonlinear ISI in direct detection long-haul fiber optic systems

Long-distance, high-rate fiber optic systems employing directly modulated 1.55- mu m single-mode lasers and conventional single-mode fiber suffer severe intersymbol interference (ISI) with a large nonlinear component. A method of reducing the nonlinearity of the ISI, thereby making linear equalization more viable, is investigated. It is shown that the degree of nonlinearity is highly dependent on the choice of laser bias current, and that in some cases the ISI nonlinearity can be significantly reduced by biasing the laser substantially above threshold. Simulation results predict that an increase in signal-to-nonlinear-distortion ratio as high as 25 dB can be achieved for synchronously spaced samples at an optimal sampling phase by increasing the bias current from 1.2 times threshold to 3.5 times threshold. The high SDR indicates that a linear tapped delay line equalizer could be used to mitigate ISI. Furthermore, the shape of the pulse response suggests that partial response precoding and digital feedback equalization would be particularly effective for this channel.<<ETX>>