This paper presents a semi-automatic algorithm for video object segmentation. Our algorithm assumes the use of multiple key video frames in which a semantic object of interest is defined in advance with human assistance. For video frames between every two key frames, the specified video object is tracked and segmented automatically using Learning Vector Quantization (LVQ). Each pixel of a video frame is represented by a 5-dimensional feature vector integrating spatial and color information. We introduce a parameter K to adjust the balance of spatial and color information. Experimental results demonstrate that the algorithm can segment the video object consistently with less than 2% average error when the object is moving at a moderate speed.

Various types of radars have been developed and used until now--such as Pulse, FM-CW, and Spread Spectrum. Additionally, we have proposed another type of radar which measures distances by using standing wave. We have named it as "Standing Wave Radar." It has a shorter minimum detectable range and higher accuracy compared to other types. However, the radar can not measure distances down to zero meters like other types of radars. Minimum detectable range of the standing wave radar depends on a usable frequency range. A wider frequency range is required if we need to measure shorter distances. Consequently, we propose a new method for measuring distances down to zero meters without expanding the frequency range. We use an analytic signal, which is a complex sinusoidal signal. The signal is obtained by observing the standing wave with multiple detectors. We calculate distances by Fourier transform of the analytic signal. Moreover, we verify the validity of our method by simulations based on numerical calculation. The results show that it is possible to measure distances down to zero meters. In our method, measurement errors are caused by deviations of position and gain of the detectors. They are around 10cm at the largest if the gain deviations are up to 1% and the position deviations are up to 6% of the spacing between the detectors. Prevalent radars still have a common defect that they can not measure distances from zero to several meters. We expect that the defect will be eliminated by putting our method into practical use.

In this letter, the problem of attaining frequency synchronization for high rate WPANs is considered. An estimation of the frequency offset is essential step at the receiver because the frequency offset means that the phase of transmitted signal is corrupted. The proposed algorithm utilizes the difference in phase between received preamble and estimated preamble. It turns out that the proposed algorithm has low complexity and good performance. To enhance the capability, we use peak phase detector. The performance of the proposed algorithm is assessed by computer simulation, and the frequency offset variance is compared to Cramer Rao Bound.

For three-step cell search in WCDMA, a stepwise serial scheme is conventionally employed, where each step of cell search operates in serial. In order to reduce the cell search time, a stepwise parallel scheme can be considered as a candidate for cell search, where each step operates in pipelined fashion. However, in the parallel scheme where the processing time in every step is equal, excessive accumulations are caused in step (1) and step (3) because the period of a code used for step (2) is much longer than that for the other steps. And it does not provide any gain because the effect becomes easily saturated with the number of accumulations. In this paper, the performance of parallel cell search is analyzed, and compared with that of serial cell search. Finally, it is shown that the performance of parallel cell search can be improved by adjusting the processing time in each step, based on the analytic results.

In this paper, a new tracking method for Code Shift Keying Spread Spectrum (CSK/SS) is proposed. In the CSK/SS systems, since the transmitted sequences vary at every frame, synchronization is difficult. In the proposed method, a pair consists of the Manchester-coded PN sequence and the non-Manchester-coded PN sequence as the synchronizing sequence. The cross-correlation characteristic of the pair is used instead of the S-curve of delay lock loop (DLL). The receiver can track the signal timing by using the characteristic. The proposed method is applied to the external synchronization system whose structure is simple. The following performance in channel with the Additive White Gaussian Noise (AWGN) is analyzed and evaluated; (1) the tracking error (jitter) performance and (2) the bit error rate (BER) performance that takes the jitter into account. As a result, the jitter of the proposed system is better than those of the conventional DLL systems as Eb/N0 increases. The jitter of the proposed system has less degradation than those of the conventional systems, even if the number of users increases. Moreover, BER of the proposed system is similar to that of the 1Δ-DLL system and superior to that of the 2Δ-DLL system.

We present a method of stabilizing a class of nonlinear systems which are not necessarily feedback linearizable. First, we show a new way of constructing a diffeomorphism to transform a class of nonlinear systems to the feedback linearized form with perturbation. Then, we propose a semi-globally stabilizing control law for nonlinear systems that are connected by a chain of integrator perturbed by arbitrary nonlinear terms. In our approach, we have flexibility in choosing a diffeomorphism where the system is not restricted to involutivity and this leads to reduction in computational burden and flexibility in controller design.

A folded loop antenna for handsets has already been introduced and shown as one of balance-fed antennas for handsets, which is very effective to mitigate the antenna performance degradation due to the body effect. In order to meet the requirements for the latest handsets such as low profile and small size, a folded loop antenna is modified. The antenna, which is possibly built in the handsets, is newly proposed. Low profile and small size is achieved by consisting of the half of low profile folded loop antenna, which has a structure folded loop elements sideways so that the antenna can be placed on the ground plane (GP). In the analysis, the electromagnetic simulator based on the FDTD (Finite Difference Time Domain) method is used and the design parameters useful in practical operation are found. The electromagnetic simulator based on the Method of Moment (MoM) is used to calculate the current distribution on the antenna element and the GP. An example of low profile and small size antenna which has wideband characteristics are designed based on these parameters, and the antenna characteristics such as VSWR, the current distributions and the radiation patterns are compared with Planar Inverted-F Antenna (PIFA), which is one of conventional built-in antennas for handset. As a result, it has been confirmed that the physical volume of the antenna, which has been introduced here, becomes smaller than that of PIFA. In addition, the radiation efficiency of these antennas is measured and the results are compared with each other.

A new approach is proposed in this paper for the tracking of the optimal operating frequency in a Class E backlight inverter using the phase-locked loop (PLL) technique. First, a new single-stage backlight module is introduced to simplify the circuit and to raise the system efficiency. A piezoelectric transformer (PT) is used to drive the cold cathode fluorescent lamp (CCFL) to eliminate the downside of a conventional transformer and to reduce the dimension of the backlight module. Next, a PLL is embedded in the backlight system, as a feedback mechanism, to track the optimal operating frequency of the PT so that the PT's temperature effect is removed and, hence, the system efficiency and stability is improved. The feedback variable proposed is a phase angle rather than a lamp current amplitude traditionally used. A simplified model, along with its design procedure, is next presented. The complete analysis and design considerations are detailed. Finally, it is rather encouraging to observe that the experimental results match our analytical solutions closely.

In many electromagnetic field problems, matrix equations were always deduced from using the method of moment. Among these matrix equations, some of them might require a large amount of computer memory storage which made them unrealistic to be solved on a personal computer. Virtually, these matrices might be too large to be solved efficiently. A fast algorithm based on a Toeplitz matrix solution was developed for solving a bordered Toeplitz matrix equation arising in electromagnetic problems applications. The developed matrix solution method can be applied to solve some electromagnetic problems having very large-scale matrices, which are deduced from the moment method procedure. In this paper, a study of a computationally efficient order-recursive algorithm for solving the linear electromagnetic problems [Z]I = V, where [Z] is a Toeplitz matrix, was presented. Upon the described Toeplitz matrix algorithm, this paper derives an efficient recursive algorithm for solving a bordered Toeplitz matrix with the matrix's major portion in the form of a Toeplitz matrix. This algorithm has remarkable advantages in reducing both the number of arithmetic operations and memory storage.

Larger object classes often become more costly classes in the maintenance phase of object-oriented software. Consequently class would have to be constructed in a medium or small size. In order to discuss such desirable size, this paper proposes a simple method for predictively discriminating costly classes in version-upgrades, using a class size metric, Stmts. Concretely, a threshold value of class size (in Stmts) is provided through empirical studies using many Java classes. The threshold value succeeded as a predictive discriminator for about 73% of the sample Java classes.

Periodically nonuniform coupled microstrip line (PNC-ML) loaded with transverse slits is characterized using the fullwave method of moments and short-open calibration technique. Guided-wave characteristics of both even- and odd-modes are thoroughly investigated in terms of two extracted per-unit-length transmission parameters, i.e., phase constants and characteristic impedances. As such, frequency-dependent coupling between the lines of the finite-extended PNCML is exposed via two dissimilar impedances. Meanwhile, two phase constants try to be equalized at a certain frequency by properly adjusting the slit depth and periodicity, aiming at realizing the transmission zero. Further, equivalent J-inverter network parameters of this finite-length PNCML are derived to reveal the relationship between the transmission zero and harmonic resonance. By allocating this zero to the frequency twice the fundamental passband, one-stage and two-stage PNCML filters are then designed, fabricated and measured to showcase the advantageous capacity of the proposed technique in harmonic suppression.

This paper presents a reduced-complexity maximum likelihood detection (MLD) scheme for orthogonal frequency division multiplexing with space division multiplexing (OFDM-SDM) systems. Original MLD is known to be an optimal scheme for detecting the spatially multiplexed signals. However, MLD suffers from an exponentially computational complexity because it involves an exhaustive search for the optimal result. In this paper, we propose a novel detection scheme, which drastically reduce the complexity of MLD while keeping performance losses small. The proposed scheme decouples the spatially multiplexed signals in two stages. In stage one, the estimated symbols obtained from zero-forcing (ZF) are used to limit the candidate symbol vectors. In stage two, to form a final estimate of the transmitted symbol vector, the Euclidean or original defined likelihood metric is examined over all symbol vectors obtained from stage 1. Both the bit error rate (BER) and packet error rate (PER) performances are evaluated over a temporally and spatially uncorrelated frequency selective channel through the computer simulations. For a four-transmit and four-receive OFDM-SDM system transmitting data at 144 Mbit/s and 216 Mbit/ss i.e., employing 16 Quadrature Amplitude Modulation (16QAM) and 64QAM subcarrier modulation over 16.6 MHz bandwidth channel, the degradation in required SNR from MLD for PER = 1% are about 0.6 dB and 1.5 dB, respectively. However, the complexity of MLD is reduced to 0.51000% and 0.01562%.

In this letter, we propose a recursive space time decoding method for orthogonal frequency division multiplexing (OFDM) systems exploiting multiple transmit antenna diversity when the channels are fast fading. We first develop a computationally efficient space-time decoding method involving a matrix inversion to mitigate the channel variation effect. We then further reduce the computational complexity of the matrix inversion decoding method via a recursive formulation. Computer simulation results show that the proposed recursive decoding has much better BER performance than Alamouti decoding, requiring much less computation than the matrix inversion decoding. Moreover, the relative advantage in BER performance of the proposed scheme over Alamouti decoding stands out as the Doppler frequency increases.

Intelligent robot control systems based on multiprocessors, sensors, and actuators require a flexible network for communicating various types of real-time data (e.g. sensing data, interrupt signals). Furthermore, serial data transfer implemented using a few wiring lines is also required. Therefore, a CDMA serial communication interface with a new two-step synchronization technique is proposed to counter these problems. The transmitter and receiver fabricated by 0.25 µm digital CMOS technology achieve 2.7 Gcps (gigachips per second) and can multiplex 7 communication channels.

Polling I/O mechanisms on the Unix platform such as select() and poll() cause high processing overhead when they are used in a heavily-loaded network server with many concurrent open sockets. Large waste of processing power incurs not only service degradation but also various troubles such as high electronic power consumption and worsened MTBF of server hosts. It is thus a serious issue especially in large-scale service providers such as an Internet data center (iDC) where a great number of heavily-loaded network servers are operated. As a solution of this problem, we propose a technique of fine-grained control on the invocation intervals of the polling I/O function. The uniqueness of this study is the utilization of POSIX real-time scheduling to enable the fine-grained execution control. Although earlier solutions such as an explicit event delivery mechanism also addressed the problem, they require major modification in the OS kernel and transition from the traditional polling I/O model to the new explicit event-notification model. On the other hand, our technique can be implemented with low cost because it just inserts a few small blocks of codes into the server program and does not require any modification in the OS kernel.

The goal of gaze detection is to locate the position (on a monitor) where a user is looking. Previous researches use one wide view camera, which can capture the user's entire face. However, the image resolution is too low with such a camera and the fine movements of user's eye cannot be exactly detected. So, we propose the new gaze detection system with dual cameras (a wide and a narrow view camera). In order to locate the user's eye position accurately, the narrow-view camera has the functionalities of auto focusing/panning/tilting based on the detected 3D eye positions from the wide view camera. In addition, we use the IR-LED illuminators for wide and narrow view camera, which can ease the detecting of facial features, pupil and iris position. To overcome the problem of specular reflection on glasses by illuminator, we use dual IR-LED illuminators for wide and narrow view camera and detect the accurate eye position, which is not hidden by the specular reflection. Experimental results show that the gaze detection error between the computed positions and the real ones is about 2.89 cm of RMS error.

Routing protocols are a critical component in IP networks. Among these, the Open Shortest Path First (OSPF) has been a widely used routing protocol in IP networks for some years. Beside dedicated hardware, a great interest on routing systems based on open software is raising among Internet Service Providers. Many open source implementations of this protocol have been developed, among which GNU Zebra is one of the most complete. In this paper we perform a study of the performances of the Shortest Path First computation in GNU Zebra, as prescribed by the Internet Engineering Task Force, and we provide a comparison between a Cisco 2621 access router and a PC-based router equipped with routing software GNU Zebra. Moreover we describe a set of modifications made on the GNU Zebra code in order to optimize some processes, whose algorithms were not efficient and whose experimental measures had showed a lack of optimization, thus finally obtaining performances better than the one measured on commercial systems.

A novel image enhancement algorithm that can efficiently detect a small target of panoramic infrared (IR) imagery is proposed. Image enhancement is the first step for detecting and recognizing a small target in the IR imagery. The essence of the proposed algorithm is to utilize the independent histogram equalization (HE) separately over two sub-images obtained by decomposing the given image through the statistical hypothesis testing (SHT). Experimental results show that the proposed algorithm has better discrimination and lower false alarm rate than the conventional algorithms.

A speaker identification system based on wavelet transform (WT) derived from codebook design using fuzzy c-mean algorithm (FCM) is proposed. We have combined FCM and the vector quantization (VQ) algorithm to avoid typical local minima for speaker data compression. Identification accuracies of 94% were achieved for 100 Mandarin speakers.

A 3-bit flash analog-to-digital converter (ADC) for electronic dispersion compensation (EDC) was developed using InP HBTs. Nyquist operation was confirmed up to 24 Gsps, which enables oversampling acquisition for 10 Gbit/s non-return-to-zero (NRZ) signals. The ADC can also be operated at up to 37 Gsps for low input frequencies. To reduce aperture jitter and achieve a wide band of over 7 GHz, an analog input signal for all pre-amplifiers and a clock signal for all latched comparators are provided as traveling waves through coplanar transmission lines. EDC was demonstrated by capturing a 10-Gbit/s pseudo-random bit stream (PRBS) with the waveform degraded by polarization-mode dispersion (PMD). By using the captured data, we confirmed that a calculation of a transversal filter mitigates PMD.