An ultra-fast optoelectronic decision circuit using resonant tunneling diodes (RTD's) and a uni-traveling-carrier photodiode (UTC-PD) is proposed. The circuit employs two cascaded RTD's for ultra-fast logic operation and one UTC-PD that offers a direct optical input interface. This novel configuration is suitable for ultra-fast decision operation. Two types of decision circuits are introduced: a positive-logic type and a negative-logic type. Operations of these circuits were simulated using SPICE with precisely investigated RTD and UTC-PD models. In terms of circuit speed, 40-Gbit/s decision and 80-Gbit/s demultiplexing were expected. Furthermore, the superiority of the negative-logic type in terms of the circuit operating margin and the relationship between input peak photocurrent and effective logic swing were clarified by SPICE simulations. In order to confirm the basic functions of the circuits and the accuracy of the simulations, circuits were fabricated by monolithically integrating InP-based RTD's and UTC-PD's. The circuits successfully exhibited 40-Gbit/s decision operation and 80-Gbit/s demultiplexing operation with less than 10-mW power dissipation. The superiority of the negative-logic type circuit for the circuit operation was confirmed, and the relationship between the input peak photocurrent and the effective logic swing was as predicted.

High-speed collision detection is important to realize a highly-safe intelligent vehicle. In collision detection, high-computational power is required to perform matching operation between discrete points on surfaces of a vehicle and obstacles in real-world environment. To achieve the highest performance, a hierarchical matching scheme is proposed based on two representations: the coarse representation and the fine representation. A vehicle is represented as a set of rectangular solids in the fine representation (fine rectangular solids), and the coarse representation, which is also a set of rectangular solids, is produced by enlarging the fine representation. If collision occurs between an obstacle discrete point and a rectangular solid in the coarse representation (coarse rectangular solid), then it is sufficient to check the only fine rectangular solids contained in the coarse one. Consequently, checks for the other fine rectangular solids can be omitted. To perform the hierarchical matching operation in parallel, a hierarchically-content-addressable memory (HCAM) is proposed. Since there is no need to perform matching operation in parallel with fine rectangular solids contained in different coarse ones, the fine ones are mapped onto a matching unit. As a result, the number of matching units can be reduced without decreasing the performance. Under the condition of the same execution time, the area of the HCAM is reduced to 46.4% in comparison with that of the conventional CAM in which the hierarchical matching scheme is not used.

This paper explores a possibility of constructing massively parallel molecular computing systems using molecular electronic devices called enzyme transistors. The enzyme transistor is, in a sense, an artificial catalyst which selects a specific substrate molecule and transforms it into a specific product. Using this primitive function, various active continuous media for signal transfer/processing can be realized. Prominent examples discussed in this paper are: (i) Turing pattern formation and (ii) excitable wave propagation in a two-dimensional enzyme transistor array. This paper demonstrates the potential of enzyme transistors for creating reaction-diffusion dynamics that performs useful computations in a massively parallel fashion.

Feature selection is an integral part of any pattern recognition system. Removal of redundant features improves the efficiency of a classifier as well as cut down the cost of future feature extraction. Recently neural network classifiers have become extremely popular compared to their counterparts from statistical theory. Some works on the use of artificial neural network as a feature selector have already been reported. In this work a simple feature selection algorithm has been proposed in which a fractal neural network, a modified version of multilayer perceptron, has been used as a feature selector. Experiments have been done with IRIS and SONAR data set by simulation. Results suggest that the algorithm with the fractal network architecture works well for removal of redundant informations as tested by classification rate. The fractal neural network takes lesser training time than the conventional multilayer perceptron for its lower connectivity while its performance is comparable to the multilayer perceptron. The ease of hardware implementation is also an attractive point in designing feature selector with fractal neural network.

The concept of a schedulable region (SR) was introduced to characterize the capacity of a multiplexer and provide a separation between call-level and cell-level phenomena. In this paper, we present a framework and algorithm for real-time estimation of the schedulable region. A major problem associated with online estimation is that the objects of measurement are not fixed in the presence of call arrivals and departures. The invariance property is exploited to carry out measurements in the presence of call arrivals and departures. By virtue of it, the equivalent bandwidth could be defined on the condition of the number of each traffic class call in progress. Another important thing we consider here is that the search algorithm to estimate the effective bandwidth should be chosen depending on the arrival statistics and QOS constraints. The algorithms presented here have been implemented on an ATM switch.

An analog vector quantization processor has been designed based on the neuron-MOS (νMOS) technology. In order to achieve a high integrating density, template information is merged into the matching cell (the absolute value circuitry) using the νMOS ROM technology. A new-architecture νMOS winner-take-all (WTA) circuit is employed for fully-parallel search for the minimum-distance vector. The WTA performs multi-resolution winner search with an automatic feedback gain control. A test chip having 256 16-element fixed template vectors has been built in a 1.5-µm double-polysilicon CMOS technology with the chip size of 7.2 mm 7.2 mm, and the basic operation of the circuits has been demonstrated.

Recent status of the polymer optical fiber (POF) for high speed data communication and telecommunication is reviewed. The GI POF was proposed for the first time 20 years ago at Keio University, and several methodologies to fabricate GI POF have been currently proposed worldwide. In this paper, we both theoretically and experimentally verify that the most transparent GI POF can be obtained by the polymer-dopant system. The relation between the refractive index profile and the dispersion characteristics of the GI POF was quantitatively clarified. The refractive index profile of the GI POF obtained by the interfacial-gel polymerization process was controlled to enable to transmit the order of gigabit per second bit rate. Furthermore, the accurate approximation of the refractive index profile and consideration of mode dependent attenuation enabled to precisely predict the dispersion characteristics of the GI POF.

Texture has been investigated as a cue for reconstructing 3-D structure. There are various textures in a natural scene. In this paper, the regularity of alignment of texture elements was manipulated to investigate its effect on human perception. The results show that the regularity affects human perception when only the texel density gradient is given as cue or the density cue is inconsistent with the compression cue. We introduce a model based on a MAP estimation to account for the result from a viewpoint of an integration of 3-D cues. The model simultaneously estimates texture properties and 3-D surface orientation by using prior knowledge about texture and 3-D surface. The performance of the model accounts for the experimental result well.

WDM transmission experiments over cascaded sections of optical links including wavelength converting 2R-transponders have been carried out in a loop testbed. Using dispersion compensated links and simple direct modulated transponder lasers, up to 11 cascaded crossconnects and 1750 km trunk lines have been bridged with 2.5 Gbit/s NRZ signals. The limitations are given mainly due to the accumulated jitter as it is shown by numerical simulations. The results indicate, that 2R-transponders are a useful approach to a flexible WDM network design using bitrate-transparent wavelength conversion.

We analyze the dispersion-managed optical transmission system for the non-return-to-zero (NRZ) pulse format. First, we investigate the physical image of dispersion management by computing small-signal-based transfer functions, and summarize the dependence of transmission performance on system parameters. Next, the Q-map is computed numerically to design long-distance large-capacity dispersion-managed transmission systems for a single channel in a more detailed manner. It is shown that the third-order dispersion of fibers negatively influences transmission performance, and third-order dispersion compensation is proved to be an effective method for extending the transmission distance of high bit-rate systems. Utilizing these results, guidelines can be derived for the optimal design of long-distance large-capacity NRZ transmission systems.

An optical add-drop multiplexer with a grating-loaded directional coupler in silica waveguides is demonstrated. The device for this configuration has a large fabrication tolerance and is small in size. A new scheme, in which the coupling length of the directional coupler is twice the complete coupling length, enables low cross-talk for both add and drop operations. This device is polarization-independent due to its relatively low-temperature process.

A 1.55-µm hybrid integrated wavelength-converter module was fabricated using a two-channel spot-size converter integrated semiconductor optical amplifier (SS-SOA) on a planar-lightwave-circuit (PLC) platform. Clear eye opening and penalty-free wavelength conversion were obtained at 2.5-Gb/s modulation with a wide wavelength difference of 46 nm. The module showed good characteristics including low insertion loss (0.1 dB), and high conversion efficiency (-0.2 dB). It also showed stable wavelength conversion for as wide as a 13 temperature range.

After a short introduction to the different requirements to and techniques for wavelength conversion, focus is on cross-gain and cross-phase modulation in SOA based converters. Aspects like jitter accumulation, regeneration and conversion to the same wavelength is discussed. It is predicted that jitter accumulation can be minimised while also assuring a high extinction ratio by using a 9-10 dB ratio between the signal and CW power. Using this guideline simulations show that 20 cross-gain modulation converters can be cascaded at 10 Gbit/s with only 20 ps of accumulated jitter and an extinction ratio of 10 dB. The regenerative capabilities of the cross-phase converters are described and verified experimentally at 20 Gbit/s. By controlling the input power to an EDFA, the noise redistribution and improvement of the signal-to-noise ratio is demonstrated. In a similar experiment at 2.5 Gbit/s, the regeneration causes a reduction of the required input power to an in-line EDFA of 6 dB for a power penalty of 1 dB at a bit error rate of 10-9. If two converters are concatenated the power requirement is reduced 8 dB. Obviously, the power reduction allows for longer spans between in-line EDFAs. A simple scheme for regeneration without wavelength conversion is assessed at 2.5 Gbit/s resulting in 4.5 dB lower required EDFA input power. The scheme is characterised by a quasi-digital transfer function that is ideal for regeneration. A combination of cross-gain and cross-phase conversion is used to perform conversion to the same wavelength at 20 Gbit/s. The insertion penalty for this dual-stage converter is below 2 dB and is mainly caused by extinction ratio degradation from the cross-gain converter. Finally, a new device for all-optical wavelength conversion has been proposed and 2.5 Gbit/s operation has been simulated with good results.

We developed an optical transceiver diode (TRAD) module for bi-directional time-compression-multiplexing (TCM) transmission systems. A wavelength-insensitive structure as a receiver and a low-capacitance configuration in the module provide a high sensitivity. Stable switching of 156 Mbit/s NRZ burst signals between the transmitter and receiver modes is achieved. In addition, it is shown that optical module cost can be further reduced by using passive alignment on a Si bench.

A novel 1470-nm-band (S+ band) wavelength-division multiplexing (WDM) transmission system is described. The first advantage of S+-band transmission is suppression of degradation caused by four-wave mixing (FWM), which has been the dominant impairment factor in WDM transmission systems on dispersion-shifted fibers (DSFs). FWM suppression by using the S+ band instead of the conventional 1550-nm-band (M band) is successfully demonstrated. The second advantage is expansion of the usable bandwidth by using the S+ band together with other wavelength bands. A triple-wavelength-band WDM repeaterless transmission experiment using the S+ band, the M band and the L band (1580-nm-band) is conducted over DSF, and it is shown that degradation due to inter-wavelength-band nonlinear interactions is negligible in the transmission. Moreover, the transmission performance of an S+-band linear repeating system is estimated by computer simulation, and compared with that of other wavelength-band systems. In the experiments, thulium-doped fiber amplifiers (TDFAs) are used for amplification of signals in the S+ band.

Aiming at wideband and flat supercontinuum generation (SC) from optical fibers in the 1.55-µm wavelength region, we study both experimentally and theoretically how SC spectra are influenced by group-velocity dispersion (GVD) of fibers. In the anomalous GVD region, since the peak power of pump pulses is kept high during propagation through the fiber by the higher-order soliton effect, the Raman effect has an adverse effect to flat and wideband SC generation. In the zero GVD region, the interplay of the third-order dispersion (TOD) and the self-phase modulation splits the SC spectrum into two main components. On the other hand, in the normal GVD region, nevertheless the SC spectrum broadens wider and smoother than those in anomalous and zero GVD regions, it is still asymmetric when TOD of the fiber can not be ignored. From these results, we find that a dispersion-flattened fiber with normal GVD is the most suitable for flat and wideband SC generation. A 280-nm wide SC spectrum with the spectral-density fluctuation less than 10 dB is actually generated from such a fiber.

Aiming at wideband and flat supercontinuum generation (SC) from optical fibers in the 1.55-µm wavelength region, we study both experimentally and theoretically how SC spectra are influenced by group-velocity dispersion (GVD) of fibers. In the anomalous GVD region, since the peak power of pump pulses is kept high during propagation through the fiber by the higher-order soliton effect, the Raman effect has an adverse effect to flat and wideband SC generation. In the zero GVD region, the interplay of the third-order dispersion (TOD) and the self-phase modulation splits the SC spectrum into two main components. On the other hand, in the normal GVD region, nevertheless the SC spectrum broadens wider and smoother than those in anomalous and zero GVD regions, it is still asymmetric when TOD of the fiber can not be ignored. From these results, we find that a dispersion-flattened fiber with normal GVD is the most suitable for flat and wideband SC generation. A 280-nm wide SC spectrum with the spectral-density fluctuation less than 10 dB is actually generated from such a fiber.

A 1.55-µm hybrid integrated wavelength-converter module was fabricated using a two-channel spot-size converter integrated semiconductor optical amplifier (SS-SOA) on a planar-lightwave-circuit (PLC) platform. Clear eye opening and penalty-free wavelength conversion were obtained at 2.5-Gb/s modulation with a wide wavelength difference of 46 nm. The module showed good characteristics including low insertion loss (0.1 dB), and high conversion efficiency (-0.2 dB). It also showed stable wavelength conversion for as wide as a 13 temperature range.

After a short introduction to the different requirements to and techniques for wavelength conversion, focus is on cross-gain and cross-phase modulation in SOA based converters. Aspects like jitter accumulation, regeneration and conversion to the same wavelength is discussed. It is predicted that jitter accumulation can be minimised while also assuring a high extinction ratio by using a 9-10 dB ratio between the signal and CW power. Using this guideline simulations show that 20 cross-gain modulation converters can be cascaded at 10 Gbit/s with only 20 ps of accumulated jitter and an extinction ratio of 10 dB. The regenerative capabilities of the cross-phase converters are described and verified experimentally at 20 Gbit/s. By controlling the input power to an EDFA, the noise redistribution and improvement of the signal-to-noise ratio is demonstrated. In a similar experiment at 2.5 Gbit/s, the regeneration causes a reduction of the required input power to an in-line EDFA of 6 dB for a power penalty of 1 dB at a bit error rate of 10-9. If two converters are concatenated the power requirement is reduced 8 dB. Obviously, the power reduction allows for longer spans between in-line EDFAs. A simple scheme for regeneration without wavelength conversion is assessed at 2.5 Gbit/s resulting in 4.5 dB lower required EDFA input power. The scheme is characterised by a quasi-digital transfer function that is ideal for regeneration. A combination of cross-gain and cross-phase conversion is used to perform conversion to the same wavelength at 20 Gbit/s. The insertion penalty for this dual-stage converter is below 2 dB and is mainly caused by extinction ratio degradation from the cross-gain converter. Finally, a new device for all-optical wavelength conversion has been proposed and 2.5 Gbit/s operation has been simulated with good results.

We developed an optical transceiver diode (TRAD) module for bi-directional time-compression-multiplexing (TCM) transmission systems. A wavelength-insensitive structure as a receiver and a low-capacitance configuration in the module provide a high sensitivity. Stable switching of 156 Mbit/s NRZ burst signals between the transmitter and receiver modes is achieved. In addition, it is shown that optical module cost can be further reduced by using passive alignment on a Si bench.