In this paper, we present a new method to enhance classification performance of a multiple classifier system by combining a boosting technique called AdaBoost.M2 and Kernel Discriminant Analysis (KDA). To reduce the dependency between classifier outputs and to speed up the learning, each classifier is trained in a different feature space, which is obtained by applying KDA to a small set of hard-to-classify training samples. The training of the system is conducted based on AdaBoost.M2, and the classifiers are implemented by Radial Basis Function networks. To perform KDA at every boosting round in a realistic time scale, a new kernel selection method based on the class separability measure is proposed. Furthermore, a new criterion of the training convergence is also proposed to acquire good classification performance with fewer boosting rounds. To evaluate the proposed method, several experiments are carried out using standard evaluation datasets. The experimental results demonstrate that the proposed method can select an optimal kernel parameter more efficiently than the conventional cross-validation method, and that the training of boosting classifiers is terminated with a fairly small number of rounds to attain good classification accuracy. For multi-class classification problems, the proposed method outperforms both Boosting Linear Discriminant Analysis (BLDA) and Radial-Basis Function Network (RBFN) with regard to the classification accuracy. On the other hand, the performance evaluation for 2-class problems shows that the advantage of the proposed BKDA against BLDA and RBFN depends on the datasets.

This paper introduces a texture analysis mechanism utilizing multiresolution technique to reduce false motion detection and hence thoroughly improve the interpolation results for high-quality deinterlacing. Conventional motion-adaptive deinterlacing algorithm selects from inter-field and intra-field interpolations according to motion. Accurate determination of motion information is essential for this purpose. Fine textures, having high local pixel variation, tend to cause false detection of motion. Based on hierarchical wavelet analysis, this algorithm provides much better perceptual visual quality and considerably higher PSNR than other motion adaptive deinterlacers as shown. In addition, a recursive 3-field motion detection algorithm is also proposed to achieve better performance than the traditional 2-field motion detection algorithm with little memory overhead.

In the current broadcasting system or Internet content distribution system, content providers distribute decoders (STB) that contain secret keys for content decryption, prior to content distribution. A content provider sends encrypted content to each user, who then decodes it with his or her STB. While users can get the services at their houses if they have an STB, it is hard for them to get the services outside their houses. A system that allowed users to carry around their secret keys would improve usability, but it would require countermeasures against secret key exposure. In this paper, we propose such an extended broadcasting system using tokens and group signature. The content providers can control the number of keys that users can use outside their houses. The system enables the broadcasters to minimize the damage caused by group signature key exposures and the user to get services outside his or her home.

In this paper, we describe a new ultrasonic oscillosensor and its application in a biological information measurement system. This ultrasonic sensor has a cylindrical tank of 26 mm (diameter)20 mm (height) filled with water and an ultrasonic probe. It detects the vibration of the target object by obtaining echo signals reflected from the water surface. This sensor can noninvasively detect the vibration of a patient by placing it under a bed frame. We propose a recognition system for humans in bed. Using this sensor, we could determine whether or not a patient is in the bed. Moreover, we propose a heart rate monitoring system using this sensor. When our system was tested on four volunteers, we successfully detected a heart rate comparable to that in the case of using an electrocardiograph. Fuzzy logic plays a primary role in the recognition. Consequently, this system can noninvasively determine whether a patient is in the bed as well as their heart rate using a constraint-free and compact device.

Full-diversity transmission for space-time block codes (STBCs) with multiple transmit antennas can be achieved by using coordinate interleaved orthogonal designs (CIODs). To effectively evaluate the performance of CIODs, we derive union upper and lower bounds on the symbol-error rate (SER) and a corresponding asymptotic diversity order of symmetric structured CIOD, in particular, with two transmit antennas over quasi-static spatially uncorrelated/correlated frequency-nonselective Rayleigh fading channels. Some numerical results are provided to verify our analysis.

To describe joint time- and frequency-selective (doubly-selective) channels in mobile broadband wireless communications, we propose to use the finite parameter model based on the same Bessel functions for each tap (Bessel model). An expression of channel estimation mean squared error (MSE) based on the finite parameter models in Orthogonal Frequency Division Multiplexing (OFDM) systems is derived. Then, our Bessel model is compared with commonly used finite parameter models in terms of the channel estimation MSE. Even if the channel taps have different channel correlations and some of the taps do not coincide with the Bessel function, the channel estimation MSE of the Bessel model is shown to be comparable or outperform existing models as validated by Monte-Carlo simulations over an ensemble of channels in typical urban and suburban environments.

This paper discusses design of Generalized Predictive Control (GPC) scheme. GPC is designed in two cases; the first is a dual-rate (DR) system, where the sampling interval of a plant output is an integer multiple of the holding interval of a control input, and the second is a fast-rate single-rate (FR-SR) system, where both the holding and sampling intervals are equal to the holding interval of the DR system. Furthermore, the relation between them is investigated, and this study gives the conditions that FR-SR and DR GPC become equivalent. To this end, a future reference trajectory of DR GPC is rewritten, and a future predictive output of the FR-SR GPC is rearranged.

Synchronization has gained attention in science recently. In electrical engineering, a power network requires generators and loads to be in synchronization. The increase in distributed dc power sources without sophisticated controls has made synchronization more difficult. This paper proposes a synchronous inverter designed for linking distributed power sources within a power network. The linkage inverters should have high confidential characteristics to keep the network stable. The frequency and phase synchronization between synchronous generators in power network has been acheived through power transmission systems. The synchronous inverters contribute in the development of the sophisticated power networks by providing distributed power sources that maintain synchronous operation.

This paper tries to answer the following question: What type of support should be given to an automobile driver when it is determined, via some method to monitor the driver's behavior and the traffic environment, that the driver's intent may not be appropriate to a traffic condition? With a medium fidelity, moving-base driving simulator, three conditions were compared: (a) Warning type support in which an auditory warning is given to the driver to enhance his/her situation recognition, (b) action type support in which an autonomous safety control action is executed to avoid an accident, and (c) the baseline condition in which no driver support is given. Results were as follows: (1) Either type of driver support was effective in accident prevention. (2) Acceptance of driver support functions varied context dependently. (3) Participants accepted a system-initiated automation invocation as long as no automation surprises were possible to occur.

Scene changes occur frequently in film broadcasting, and tend to destabilize the performance with blurred, jagged, and artifacts effects when de-interlacing methods are utilized. This paper presents an efficient VLSI architecture of video de-interlacing with considering scene change to improve the quality of video results. This de-interlacing architecture contains three main parts. The first is scene change detection, which is designed based on examining the absolute pixel difference value of two adjacent even or odd fields. The second is background index mechanism for classifying motion and non-motion pixels of input field. The third component, spatial-temporal edge-based median filter, is used to deal with the interpolation for those motion pixels. Comparing with the existed de-interlacing approaches, our architecture design can significantly ameliorate the PSNRs of the video sequences with various scene changes; for other situations, it also maintains better performances. The proposed architecture has been implemented as a VLSI chip based on UMC 0.18-µm CMOS technology process. The total gate count is 30114 and its layout area is about 710 710-µm. The power consumption is 39.78 mW at working frequency 128.2 MHz, which is able to process de-interlacing for HDTV in real-time.

In order to obtain better learning results in supervised learning, it is important to choose model parameters appropriately. Model selection is usually carried out by preparing a finite set of model candidates, estimating a generalization error for each candidate, and choosing the best one from the candidates. If the number of candidates is increased in this procedure, the optimization quality may be improved. However, this in turn increases the computational cost. In this paper, we focus on a generalization error estimator called the regularized subspace information criterion and derive an analytic form of the optimal model parameter over a set of infinitely many model candidates. This allows us to maximize the optimization quality while the computational cost is kept moderate.

In order to find paths guaranteed by Quality of Service (QoS), the link state database (LSDB), containing QoS constraint information, and residing in routers, needs to be well managed. However, there is a trade-off between the exact reflection of the current link status and the update cost to calculate and maintain this data. In order to perfectly reflect the current link state, each router immediately notifies its neighbors whenever link state information changes. However, this may degrade the performance of the router. On the other hand, if current link state information is not updated routinely, route setup requests may be rejected because of the discrepancy between the current link state information and the previously updated link state information in the LSDB. Therefore, we need link state update (LSU) algorithms making it possible to appropriately update the LSDB. In addition, to facilitate implementation, they also should have low-complexity and must be adaptive under the variation of network conditions. In this paper, we propose an enhanced simple-adaptive (ESA) LSU algorithm, to reduce the generation of LSU messages while maintaining simplicity and adaptivity. The performance of this algorithm is compared with five existing algorithms by rigorous simulations. The comparision shows that the ESU algorithm can adapt to changes in network conditions and its performance is superior to existing LSU algorithms.

We have been developing a secure and reliable distributed storage system, which uses a secret sharing scheme. In order to efficiently store data in the system, this paper introduces an optimal share transfer problem, and proves it to be, generally, NP-hard. It is also shown that the problem can be resolved into a Steiner tree problem. Finally, through computational experiments we perform the comparison of heuristic algorithms for the Steiner tree problem.

In this letter, we consider a class of approximately feedback linearized systems that contain both triangular and feedforward forms. With a utilization of the transformation scaling factor, we analytically show that the considered system can be globally exponentially stabilized, globally bounded, or locally stabilized depending on the shapes of triangular and feedforward forms. Our new method broadens a class of nonlinear systems under consideration over the existing results.

This paper investigates two issues of cellular engineering for cellular systems enhanced with two-hop fixed relay nodes (FRNs): spectrum partitioning and relay positioning, under the assumption of frequency reuse distance being equal to one. A channel-dependent spectrum partitioning scheme is proposed. According to this scheme, the ensemble mean of signal-to-interference-ratio on respective sets of links are taken into account to determine the bandwidths assigned to links connecting base station (BS) and FRNs, those connecting FRNs and mobile terminals (MTs) and those connecting BS and MTs. The proper FRN positioning is formulated as a constraint optimization problem, which tries to maximize the mean user data rate while at the same time ensures in probability 95% users being better served than in conventional cellular systems without relaying. It is demonstrated with computer simulations that FRN positioning has a strong impact on system performance. In addition, when FRNs can communicate with BS over line-of-sight channels the FRN enhanced cellular system with our proposed spectrum partitioning can remarkably outperform that with a known channel-borrowing based scheme and the conventional cellular systems without relaying. Simulation results also show that with proper FRN positioning the proposed spectrum partitioning scheme is robust against the unreliability of links connecting BS and FRNs.

This paper analyses the autocorrelation function of return waveforms in high precision radar altimeters employing chirp-pulse transmit signal under the condition of near-nadir deviations of the antenna boresight axis. It is shown that in case of ultra wideband transmit signals providing very high time resolution the correlation function can be approximated by a product of two separate functions of time.

In a frequency-selective multiple-input multiple-output (MIMO) channel, the optimum transmission is achieved by beamforming with eigenvectors obtained at each discrete frequency point, i.e., an extension of eigenbeam-space division multiplexing (E-SDM). However, the calculation load of eigenvalue decomposition at the transmitter increases in proportion to the number of frequency points. In addition, frequency-independent eigenvectors increase the delay spread of the effective channel observed at the receiver. In this paper, we propose a pseudo eigenvector scheme for the purpose of mitigating the calculation load and maintaining frequency continuity (or decreasing the delay spread). First, we demonstrate that pseudo eigenvectors reduce the delay spread of the effective channels with low computational complexity. Next, the practical performance of the pseudo E-SDM (PE-SDM) transmission is evaluated. The simulation results show that PE-SDM provides almost the same or better performance compared with E-SDM when the receiver employs a time-windowing-based channel estimation available in the low delay spread cases.

In this letter, we introduce a new adaptive beamforming assisted receiver based on minimizing the approximate negentropy (NEGMIN) of the estimation error. We consider the approximate negentropy by using a nonpolynomial expansion of the estimation error as a new performance criterion to improve the performance of adaptive beamforming receivers based on the minimizing minimum mean squared error (MMSE) criteria. Negentropy includes higher order statistical information and its minimization provides improved converge, performance, and accuracy compared to traditional methods such as MMSE in terms of bit error rate (BER).

This paper addresses the ambiguity of radar altimetry related to the statistical nature of the Earth's surface roughness. A hypothetic altimetry method, which provides a simple way to quantify the measures of ambiguity, is proposed. Cramer-Rao lower bounds on the variances of estimates for the mean altitude and root-mean-square height of the sea surface are suggested as such quantitative measures. The accuracy of some types of air- and spaceborne radar altimeters is numerically analyzed against the derived lower bounds.

The open-vocabulary name recognition technique is one of the most challenging tasks in the application of automatic Chinese speech recognition technology. It can be used as the free name input method for telephony speech applications and automatic directory assistance systems. A Chinese name usually has two to three characters, each of which is pronounced as a single tonal syllable. Obviously, it is very confusing to recognize a three-syllable word from millions to billions of possible candidates. A novel interactive automatic-speech-recognition system is proposed to resolve this highly challenging task. This system was built as an open-vocabulary Chinese name recognition system using character-based approaches. Two important character-input speech-recognition modules were designed as backoff approaches in this system to complete the name input or to correct any misrecognized characters. Finite-state networks were compiled from regular grammar of syllable spellings and character descriptions for these two speech recognition modules. The possible candidate names cover more than five billions. This system has been tested publicly and proved a robust way to interact with the speaker. An 86.7% name recognition success rate was achieved by the interactive open-vocabulary Chinese name input system.