The issue of scalable Differentiated Services (DiffServ) admission control now is still an open research problem. We propose a new admission control model that can not only provide coarse grain Quality of Services (QoS), but also guarantee end-to-end QoS for assured service without per-flow state management at core routers within DiffServ domain. Associated with flow aggregation model, a hybrid signaling protocol is proposed to select the route satisfying the end-to-end QoS requirements. Simulation result shows that the proposed model can accurately manage resource, leading to much better performance when compared to other schemes.

Finding DC operating points of transistor circuits is an important and difficult task. The Newton-Raphson method adopted in SPICE-like simulators often fails to converge to a solution. To overcome this convergence problem, homotopy methods have been studied from various viewpoints. For efficiency of globally convergent homotopy methods, it is important to give an appropriate initial solution as a starting point. However, there are few studies concerning such initial solution algorithms. In this paper, initial solution problems in homotopy methods are discussed, and an effective initial solution algorithm is proposed for globally convergent homotopy methods, which finds DC operating points of transistor circuits efficiently. Numerical examples using practical transistor circuits show the effectiveness of the proposed algorithm.

In this letter, a receive frequency diversity technique is proposed to improve the performance of a multiplexed STBC OFDM system. Frequency diversity in the multiplexed STBC OFDM system is obtained by introducing frequency shifter in the successive STBC symbols and applying MRRC technique to regenerated and subtracted signals of the predecoded data from multiplexed STBC decoder. It is shown by computer simulation that the performance of the proposed multiplexed STBC OFDM systems with frequency diversity is improved by 5 dB at the BER of 10-3 over the existing multiplexed STBC OFDM systems with the same data rate.

In this letter, we propose a groupwise successive interference cancellation (GSIC) receiver with gradient descent search for multi-rate DS-CDMA system. Proposed receiver incorporates iterative gradient descent search algorithm into conventional GSIC receiver for multi-rate DS-CDMA system. It is shown that the receiver achieves significant performance improvement over the matched filter (MF) receiver, GSIC receiver, multi-stage parallel interference cacnellation (PIC) receiver, multi-stage partial PIC receiver, and GSIC receiver with PIC in a Rayleigh fading channel.

The classification time required by conventional multi-class SVMs greatly increases as the number of pattern classes increases. This is due to the fact that the needed set of binary class SVMs gets quite large. In this paper, we propose a method to reduce the number of classes by using nearest neighbor rule (NNR) in the principle component analysis and linear discriminant analysis (PCA+LDA) feature subspace. The proposed method reduces the number of face classes by selecting a few classes closest to the test data projected in the PCA+LDA feature subspace. Results of experiment show that our proposed method has a lower error rate than nearest neighbor classification (NNC) method. Though our error rate is comparable to the conventional multi-class SVMs, the classification process of our method is much faster.

In order to benefit from the advantages of soft handoff (SHO), it is important that the SHO parameters (the SHO thresholds; T_ADD and T_DROP are well assigned. T_ADD is the threshold used for triggering a pilot with high strength to be added to the Active Set (AS) list. The AS means the pilots associated with the forward traffic channels assigned to mobile station. In contrast, T_DROP is the threshold used for triggering a pilot with low strength to be dropped from the AS list. This paper analyzes the effects of varying SHO thresholds in a cellular code division multiple access (CDMA) system on the blocking probability based on traffic load and geometrical distances in hexagonal layout of base stations (BSs). In addition, the previously proposed traffic load equation is applied to the proposed SHO model for balancing the numbers of new and handoff calls on the forward link capacity in case of uniform traffic load. The results show that the blocking probability is more sensitive to T_DROP than to T_ADD variations.

Flexible Resource Allocation (FRA) strategies selectively control the transmission rates of users allowing them to specify maximum and minimum bandwidth requirements for the service type requested ensuring a minimum quality of service (QoS) is met. Complete, Partial, and Non Resource Sharing are the three types of resource sharing policies that can be used in systems with integrated services (voice, video and data) with different QoS and elasticities requirements. In this paper, an FRA strategy with Partial Resource Sharing, called Primary Unavailable Secondary Minimum (PUSMin), is presented. An analytical method is developed to assess its performance in an environment where several service types (with different bandwidth and elasticities requirements) exist. Results show that PUSMin decreases the resource reassignment rate as the offered traffic increases. This decreases the signalling overhead and computational complexity in the Base Station Controller (BSC) or Base Transceiver Station (BTS).

In this letter, the concept of cross-entropy is introduced for unsupervised polarimetric synthetic aperture radar (SAR) image classification. The difference between two scatterers is decomposed into three parts, i.e., the difference of average scattering characteristic, the difference of scattering randomness and the difference of scattering matrix span. All these three parts are expressed in cross-entropy formats. The minimum cross-entropy principle is adopted to make classification decision. It works well in unsupervised terrain classification with a NASA/JPL AIRSAR image.

Classical studies of Asynchronous Transfer Mode (ATM) switching elements and in particular the buffer behavior of the Shared Buffer Memory (SBM), assume that all read and write operations of cells to, respectively from, the SBM are executed simultaneously. However, in a real switching element, the inlets (outlets) are scanned sequentially for arriving (departing) cells during the so-called input (output) cycle. Furthermore, the input and output cycles are intermingled, each read operation being followed by a write operation. This is referred to as the Timeslot Interchange Mechanism (TIM). In this paper, we present the analysis of a queueing model that includes the TIM. We model the cell arrival processes on the inlets of the switching element as independent Bernoulli arrival processes. Moreover, we assume that cells are routed from the inlets to the outlets of the switching element according to an independent and uniform process, i.e., the destinations of consecutive cell arrivals on any given inlet are independent and for a given cell all destinations are equiprobable. Under these assumptions, we will derive expressions for the probability generating functions of the queue length in an individual routing group (a logical queue that contains all cells scheduled for the same destination), the (total) queue length in the SBM, and the cell waiting time. From these results, expressions for the mean values and the tail distributions of these quantities are calculated, and the influence of the TIM on the buffer behavior is studied through comparison with a model where all read and write operations occur simultaneously.

In this paper, a novel Wold decomposition algorithm is proposed to address the issue of deterministic component extraction for texture images. This algorithm exploits the wavelet-based singularity detection theory to process both harmonic a nd evanescent features from frequency domain. This exploitation is based on the 2D Lebesgue decomposition theory. When applying multiresolution analysis techniq ue to the power spectrum density (PSD) of a regular homogeneous random field, its indeterministic component will be effectively smoothed, and its deterministic component will remain dominant at coarse scale. By means of propagating these positions to the finest scale, the deterministic component can be properly extracted. From experiment, the proposed algorithm can obtain results that satisfactorily ensure its robustness and efficiency.

A major challenge in the design of multimedia networks is to meet the quality of service (QoS) requirements of all admitted users. Regulation and scheduling are key factors for fulfilling such requirements. We propose a rate-based regulation-scheduling scheme in which the regulation function is modulated by both the tagged stream's characteristics and the state information fed-back from the scheduler. The rate-jitter and bandwidth share of each tagged connection are controlled appropriately by considering the system time and the queue length of the scheduler. Simulation results have shown that the proposed scheme works better than other rate-based disciplines.

In this paper, the space time transmit diversity (STTD) decoding combined with minimum mean square error (MMSE) equalization is presented for MC-CDMA downlink and uplink in the presence of multiple receive antennas. The equalization weights that minimize the MSE for each subcarrier are derived. From computer simulation, it was found that the BER performance of STTD decoding combined with MMSE equalization and Mr-antenna diversity reception using the weights derived in this paper provides the same diversity order as 2Mr-antenna receive diversity with MMSE equalization but with 3 dB performance penalty and is always better than that with no diversity. The uplink BER performance can also be improved with STTD, but the error floor still exists. However, with 2-receive antennas in addition to 2-antenna STTD, the BER floor can be reduced to around 10-5 even for the uplink.

This paper presents a technique for improving the SNR and resolution of complex bandpass ΔΣADCs which are used for wireless communication systems such as cellular phone, wireless LAN and Bluetooth. Oversampling and noise-shaping are used to achieve high accuracy of a ΔΣAD modulator. However when a multi-bit internal DAC is used inside a modulator, nonlinearities of the DAC are not noise-shaped and the SNR of the ΔΣADC degrades. For the conversion of complex intermediate frequency (IF) input signals, a complex bandpass ΔΣAD modulator can provide superior performance to a pair of real bandpass ΔΣAD modulators of the same order. This paper proposes a new noise-shaping algorithm--implemented by adding simple digital circuitry--to reduce the effects of nonlinearities in multi-bit DACs of complex bandpass ΔΣAD modulators. We have performed simulation with MATLAB to verify the effectiveness of the algorithm, and the results show that the proposed algorithm can improve the SNR of a complex bandpass ΔΣADC with nonlinear internal multi-bit DACs.

In texture mapping, anisotropic filtering methods, which require more texels, have been proposed for high-quality images. Memory bandwidth, however, is still limited by a bottleneck in the texture-filtering hardware. In this paper, we propose anisotropic texture filtering based on edge function. In generating the weight that plays a key role in filtering texels loaded from memory, the edge function gives accurate contribution of texels to the pixel intensity. The quality of images is superior to other methods. For images of the same quality, our method requires less than half the texels of other methods. In other words, the improvement in performance is more than twice that of other methods.

This paper proposes a new steganographic method utilizing features of JPEG compression. The method embeds secret information using the number of zeroes in a block of quantized DCT coefficients in minimum coding units (MCU) of JPEG images. In the method, we can embed secret information into JPEG images with degradation like that by JPEG compression. Furthermore, the method causes little change of the histogram of quantized DCT coefficients, so it is hard to perceive secret information embedded by the method. The method mainly modifies boundaries between zero and non-zero DCT coefficients, so we can use the low frequency side of DCT coefficients for another steganographic method.

We present a single-chip MPEG-4 audiovisual LSI in a 0.13 µm CMOS, 5-layer metal technology with 16 Mbit embedded DRAM, which integrates four 16 bit RISC and dedicated hardware accelerators including a 5 GOPS post filtering unit. It consumes 160 mW at 125 MHz and dissipates 80 nA in the standby mode. The chip is the world first LSI handling MPEG-4 CIF video encoding at 15 frames/sec and audio/speech encoding simultaneously.

Color histograms are effective for representing color visual features. However, the high dimensionality of feature vectors results in high computational cost. Several transformations, including singular value decomposition (SVD) and principal component analysis (PCA), have been proposed to reduce the dimensionality. In PCA, the dimensionality reduction is achieved by projecting the data to a subspace which contains most of the variance. As a common observation, the PCA basis function with the lowest frquency accounts for the highest variance. Therefore, the PCA subspace may not be the optimal one to represent the intrinsic features of data. In this paper, we apply independent component analysis (ICA) to extract the features in color histograms. PCA is applied to reduce the dimensionality and then ICA is performed on the low-dimensional PCA subspace. The experimental results show that the proposed method (1) significantly reduces the feature dimensions compared with the original color histograms and (2) outperforms other dimension reduction techniques, namely the method based on SVD of quadratic matrix and PCA, in terms of retrieval accuracy.

This paper presents QoS control enhanced architecture for VoIP networks. In this architecture we use both the probe flow delay and average loss rate measurement systems. First we apply the probability-based EMBAC scheme on our delay system. Then we propose a new probability-based EMBAC with a severe congestion consideration scheme to improve the admission control scheme in both measurement systems. We compare the performance of the enhanced systems in terms of blocking probability under the same condition of achieving average packet loss rate no greater than the certain target by setting an appropriate admission threshold in each system under each scenario. In this study, it is shown through simulations that for the same target voice average loss rate, the enhanced systems proposed in this paper outperform the conventional schemes in handling the network resources. Then we will seek to prove that, for extra traffic loads within a busy period of time and with an optimal admission threshold chosen in advance, the enhanced systems can be a powerful and reliable EMBAC tool for VoIP networks in achieving high network performance with minimum blocking probability and minimum average loss rates. Finally it is shown that the enhanced systems have reasonable scalability.

We apply a selective-sets resizable cache and a complete hierarchy SRAM for the high-performance and low-power RISC CPU core. The selective-sets resizable cache can change the cache memory size by varying the number of cache sets. It reduces the leakage current by 23% with slight degradation of the worst case operating speed from 213 MHz to 210 MHz. The complete hierarchy SRAM enables the partial swing operation not only in the bit lines, but also in the global signal lines. It reduces the current consumption of the memory by 4.6%, and attains the high-speed access of 1.4 ns in the typical case.

A new speaker feature extracted from multi-wavelet decomposition for speaker recognition is described. The multi-wavelet decomposition is a multi-scale representation of the covariance matrix. We have combined wavelet transform and the multi-resolution singular value algorithm to decompose eigenvector for speaker feature extraction not at the square matrix. Our results have shown that this multi-wavelet feature introduced better performance than the cepstrum and Δ-cepstrum with respect to the percentages of recognition.