This paper describes the recent advances in semiconductor photodiodes for use in ultra-high-speed optical systems. We developed two types of waveguide photodiodes (WG-PD) -- an evanescently coupled waveguide photodiode (EC-WG-PD) and a separated-absorption-and-multiplication waveguide avalanche photodiode (WG-APD). The EC-WG-PD is very robust under high optical input operation because of its distribution of photo current density along the light propagation. The EC-WG-PD simultaneously exhibited a high external quantum efficiency of 70% for both 1310 and 1550 nm, and a wide bandwidth of more than 40 GHz. The WG-APD, on the other hand, has a wide bandwidth of 36.5 GHz and a gain-bandwidth product of 170 GHz as a result of its small waveguide mesa structure and a thin multiplication layer. Record high receiver sensitivity of -19.6 dBm at 40 Gbps was achieved. Additionally, a monolithically integrated dual EC-WG-PD for differential phase shift-keying (DPSK) systems was developed. Each PD has equivalent characteristics with 3-dB-down bandwidth of more than 40 GHz and external quantum efficiency of 70% at 1550 nm.

Logical operations in CMOS digital integration are highly prone to fail as the amount of power supply (PS) drop approaches to failure threshold. PS voltage variation is characterized by built-in noise monitors in a 32-bit microprocessor of 90-nm CMOS technology, and related with operation failures by instruction-level programming for logical failure analysis. Combination of voltage drop size and activated logic path determines failure sensitivity and class of failures. Experimental observation as well as simplified simulation is applied for the detailed understanding of the impact of PS noise on logical operations of digital integrated circuits.

We present a novel model, named Category Constraint-Latent Dirichlet Allocation (CC-LDA), to learn and recognize natural scene category. Previous work had to resort to additional classifier after obtaining image topic representation. Our model puts the category information in topic inference, so every category is represented in a different topics simplex and topic size, which is consistent with human cognitive habit. The significant feature in our model is that it can do discrimination without combined additional classifier, during the same time of getting topic representation. We investigate the classification performance with variable scene category tasks. The experiments have demonstrated that our learning model can get better performance with less training data.

This paper addresses the problem of joint transceiver design for Tomlinson-Harashima Precoding (THP) in the multiuser multiple-input-multiple-output (MIMO) downlink under both perfect and imperfect channel state information at the transmitter (CSIT). For the case of perfect CSIT, we differ from the previous work by performing stream-wise (both inter-user and intra-user) interference pre-cancelation at the transmitter. A minimum total mean square error (MT-MSE) criterion is used to formulate our optimization problem. By some convex analysis of the problem, we obtain the necessary conditions for the optimal solution. An iterative algorithm is proposed to handle this problem and its convergence is proved. Then we extend our designed algorithm to the robust version by minimizing the conditional expectation of the T-MSE under imperfect CSIT. Simulation results are given to verify the efficacy of our proposed schemes and to show their superiorities over existing MMSE-based THP schemes.

The characteristics of violet-sensitive organic photodetectors (OPDs) utilizing polyalkylfluorene and triplet materials have been studied as a host and a dopant material, respectively. For the photo absorption layer, poly(9,9-dioctylfluorene) [PFO] and a phosphorescent iridium complex (Iridium (III) bis(2-(4,6-difluorophenyl)pyridinato-N,C2) [FIrpic] or Iridium (III) bis(2-(2'-benzothienyl)pyridinato-N,C3')(acetyl-acetonate) [(btp)2Ir(acac)]) were used as a host and a dopant material, respectively. PFO: (btp)2Ir(acac) device showed less photocurrent than PFO device because (btp)2Ir(acac) enhances recombination of the photo generated carriers in the photo absorption layer. On the other hand, PFO : FIrpic device showed larger photocurrent than PFO device due to triplet energy transfer from FIrpic to PFO. A cutoff frequency of 20 MHz was observed using a sinusoidal modulated violet laser light illumination under the reverse bias of 8 V.

One-dimensional photonic crystal (PC) with alternating layers of TiO2 and SiO2 was fabricated with spin coating and low temperature baking, resulting in a successful tuning of the PC stop band so as to block the amplified spontaneous emission (ASE) of a π-conjugate polymer film. Single PC as a substrate, not a cavity with two PC's, of the polymer film was sufficient to shift the tangential ASE to the energy at PC stop band edge, indicating that the tangential ASE propagating along the interface was modulated by its evanescent-field tail in the PC, which opens the new pathway for low-threshold coherent luminescence from an ultrathin π-conjugate polymer film with ultimate mode volume.

In this paper, a new methodology to estimate the number of competing stations in an IEEE 802.11 network, is proposed. Due to the nonlinear nature of the measurement model, an iterative nonlinear filtering algorithm, called the Scaled Unscented Filter (SUF), is employed. The SUF can provide a superior alternative to nonlinear filtering than the conventional Extended Kalman Filter (EKF), since it avoids errors associated with linearization. This approach demonstrates both high accuracy in addition to prompt reactivity to changes in the network occupancy status. In particular, the proposed algorithm shows superior performance in non saturated conditions when compared to the EKF. Numerical results demonstrate that the proposed algorithm provides a more viable method for estimation of the number of competing stations in an IEEE 802.11 network, than estimators based on the EKF.

This article presents an indoor positioning and communication platform, using fluorescent lights. We set up a practical implementation of a VLC (Visible Light Communication) system in a University building. To finalize this work, it is important that we analyze the properties of the reception signal, especially the length of the data string that can be received at different walking speed. In this paper, we present a model and a series of formulae for analyzing the relationship between positioning signal availability and other important parameters, such as sensor angle, walking speed, data transmission rate, etc. We report a series of real-life experiments using VLC system and compare the results with those generated by the formula. The outcome is an improved design for determination of the reception area with more than 97% accurate signals, and an optimal transmission data length, and transmission rate.

This letter addresses the problem of robust transceiver design for the multiuser multiple-input-multiple-output (MIMO) downlink where the channel state information at the base station (BS) is imperfect. A stochastic approach which minimizes the expectation of the total mean square error (MSE) of the downlink conditioned on the channel estimates under a total transmit power constraint is adopted. The iterative algorithm reported in [2] is improved to handle the proposed robust optimization problem. Simulation results show that our proposed robust scheme effectively reduces the performance loss due to channel uncertainties and outperforms existing methods, especially when the channel errors of the users are different.

In this paper, we propose a new mask estimation method for the computational auditory scene analysis (CASA) of speech using two microphones. The proposed method is based on a hidden Markov model (HMM) in order to incorporate an observation that the mask information should be correlated over contiguous analysis frames. In other words, HMM is used to estimate the mask information represented as the interaural time difference (ITD) and the interaural level difference (ILD) of two channel signals, and the estimated mask information is finally employed in the separation of desired speech from noisy speech. To show the effectiveness of the proposed mask estimation, we then compare the performance of the proposed method with that of a Gaussian kernel-based estimation method in terms of the performance of speech recognition. As a result, the proposed HMM-based mask estimation method provided an average word error rate reduction of 61.4% when compared with the Gaussian kernel-based mask estimation method.

We propose here a fully Spice-oriented design algorithm of op-amps for attaining the maximum gains under low power consumptions and assigned slew-rates. Our optimization algorithm is based on a well-known steepest descent method combining with nonlinear programming. The algorithm is realized by equivalent RC circuits with ABMs (analog behavior models) of Spice. The gradient direction is decided by the analysis of sensitivity circuits. The optimum parameters can be found at the equilibrium point in the transient response of the RC circuit. Although the optimization time is much faster than the other design tools, the results might be rough because of the simple transistor models. If much better parameter values are required, they can be improved with Spice simulator and/or other tools.

This paper proposes the Extended Bag-of-Visterms (EBOV) to represent semantic scenes. In previous methods, most representations are bag-of-visterms (BOV), where visterms referred to the quantized local texture information. Our new representation is built by introducing global texture information to extend standard bag-of-visterms. In particular we apply the adaptive weight to fuse the local and global information together in order to provide a better visterm representation. Given these representations, scene classification can be performed by pLSA (probabilistic Latent Semantic Analysis) model. The experiment results show that the appropriate use of global information improves the performance of scene classification, as compared with BOV representation that only takes the local information into account.

Two variable gain amplifiers (VGAs) that adopt new approximated exponential equations are proposed in this paper. The dB-linear range of the proposed VGAs is extended more than what the approximated exponential equations predict by a bias circuit technique that adopts negative feedback. The proposed VGAs feature wide gain variation, low-power, high linearity, wide control signal range, and small chip size. One of the proposed VGAs is fabricated in 0.18 µm CMOS technology and measurements show a gain variation of 83 dB (-3647 dB) with a gain error of less than 2 dB, and P1 dB/IIP3 from -55/8 to -20/20.5 dBm, while consuming an average current of 3.4 mA from a 1.8 V supply; the chip occupies 0.4 mm2. The other VGA is simulated in 0.18 µm CMOS technology and simulations show a gain variation of 91 dB (-4150 dB), and P1 dB/IIP3 from -50/-25 to -33/0 dBm, while consuming an average current of 1.5 mA from a 1.8 V supply.

An all CMOS variable gain amplifier (VGA) which features wide dB-linear gain range per stage (45 dB), low power consumption (1.32 mW), small chip size (0.13 mm2), and low supply voltage (1.2 V) is described. The dB-linear range is extended by reducing the supply voltage of the conventional V-to-I converter. The two-stage VGA implemented in 0.18 µm CMOS offers 90 dB of gain variation, 3 dB bandwidth of greater than 21 MHz, and max/min input IP3 and P1 dB, respectively, of -5/-42 and -12/-50 dBm.

In this paper, characteristics of the 2-bit recessed channel memory with lifted-charge trapping nodes are investigated. The length between the charge trapping nodes through channel, which is defined as the effective memory node length (Meff), is extended by lifting up them. The dependence of VTH window and short channel effect (SCE) on the recessed depth is analyzed. Improvement of short channel effect is achieved because the recessed channel structure increases the effective channel length (Leff). Moreover, this device shows highly scalable memory characteristics without suffering from the bottom-side effect (BSE).

This paper proposes a method to generate exceptional scenarios from a normal scenario written with a scenario language. This method includes (1) generation of exceptional plans and (2) generation of exceptional scenario by a user's selection of these plans. The proposed method enables users to decrease the omission of the possible exceptional scenarios in the early stages of development. The method will be illustrated with some examples.

The historical review of Taiwan's researching activities on the features of PECVD grown SiOx are also included to realize the performance of Si nanocrystal based MOSLED made by such a Si-rich SiOx film with embedded Si nanocrystals on conventional Si substrate. A surface nano-roughened Si substrate with interfacial Si nano-pyramids at SiOx/Si interface are also reviewed, which provide the capabilities of enhancing the surface roughness induced total-internal-reflection relaxation and the Fowler-Nordheim tunneling based carrier injection. These structures enable the light emission and extraction from a metal-SiOx-Si MOSLED.

In this paper, we propose a method that recovers a smooth high-resolution image from several blurred and roughly quantized low-resolution images. For compensation of the quantization effect we introduce measurements of smoothness, Huber function that is originally used for suppression of block noises in a JPEG compressed image [Schultz & Stevenson '94] and a smoothed version of total variation. With a simple operator that approximates the convex projection onto constraint set defined for each quantized image [Hasegawa et al. '05], we propose a method that minimizes these cost functions, which are smooth convex functions, over the intersection of all constraint sets, i.e. the set of all images satisfying all quantization constraints simultaneously, by using hybrid steepest descent method [Yamada & Ogura '04]. Finally in the numerical example we compare images derived by the proposed method, Projections Onto Convex Sets (POCS) based conventinal method, and generalized proposed method minimizing energy of output of Laplacian.

A single-chip ubiquitous sensor network (USN) system-on-a-chip (SoC) for small program memory size and low power has been proposed and integrated in a 0.18-µm CMOS technology. Proposed single-chip USN SoC is mainly consists of radio for 868/915 MHz, analog building block, complete digital baseband physical layer (PHY) and media access control (MAC) functions. The transceiver's analog building block includes a low-noise amplifier, mixer, channel filter, receiver signal-strength indication, frequency synthesizer, voltage-controlled oscillator, and power amplifier. In addition, digital building block consists of differential binary phase-shift keying (DPSK) modulation, demodulation, carrier frequency offset compensation, auto-gain control, embedded 8-bit microcontroller, and digital MAC function. Digital MAC function supports 128 bit advanced encryption standard (AES), cyclic redundancy check (CRC), inter-symbol timing check, MAC frame control, and automatic retransmission. These digital MAC functions reduce the processing power requirements of embedded microcontroller and program memory size by up to 56%. The cascaded noise figure and sensitivity of the overall receiver are 9.5 dB and -99 dBm, respectively. The overall transmitter achieves less than 6.3% error vector magnitude (EVM). The current consumption is 14 mA for reception mode and 16 mA for transmission mode.

Owing to simple structure, low cost and high performance, NRD-guide millimeter wave circuits have attracted much attention in recent years. In this paper, a variety of NRD-guide passive components are reviewed with emphasis on design techniques and performance estimation in the 60 GHz frequency band where the license-free advantage is available. The passive components to be discussed here include compact bends, wideband hybrid couplers, practical three-port junctions, versatile E-plane filters, and effective feeding structures for lens antennas. Some of them are employed to construct millimeter wave transceivers. Eye patterns observed at 1.5 Gbps confirm the potential ability of the fabricated NRD-guide transceivers for high bit-rate, wireless applications.