A PAC-Bayesian margin bound for linear classifiers

We present a bound on the generalization error of linear classifiers in terms of a refined margin quantity on the training sample. The result is obtained in a probably approximately correct (PAC)-Bayesian framework and is based on geometrical arguments in the space of linear classifiers. The new bound constitutes an exponential improvement of the so far tightest margin bound, which was developed in the luckiness framework, and scales logarithmically in the inverse margin. Even in the case of less training examples than input dimensions sufficiently large margins lead to nontrivial bound values and-for maximum margins-to a vanishing complexity term. In contrast to previous results, however, the new bound does depend on the dimensionality of feature space. The analysis shows that the classical margin is too coarse a measure for the essential quantity that controls the generalization error: the fraction of hypothesis space consistent with the training sample. The practical relevance of the result lies in the fact that the well-known support vector machine is optimal with respect to the new bound only if the feature vectors in the training sample are all of the same length. As a consequence, we recommend to use support vector machines (SVMs) on normalized feature vectors only. Numerical simulations support this recommendation and demonstrate that the new error bound can be used for the purpose of model selection.     

Soft trellis-based decoder for linear block codes

A systematic design of a trellis-based maximum-likelihood soft-decision decoder for linear block codes is presented. The essence of the decoder is to apply an efficient search algorithm for the error pattern on a reduced trellis representation of a certain coset. Rather than other efficient decoding algorithms, the proposed decoder is systematically designed for long codes, as well as for short codes. Computational gain of up to 6 is achieved for long high-rate codes over the well-known trellis decoder of Wolf (1978). Efficient decoders are also obtained for short and moderate length codes     

An efficient maximum-likelihood-decoding algorithm for linear blockcodes with algebraic decoder

A new soft decoding algorithm for linear block codes is proposed. The decoding algorithm works with any algebraic decoder and its performance is strictly the same as that of maximum-likelihood-decoding (MLD). Since our decoding algorithm generates sets of different candidate codewords corresponding to the received sequence, its decoding complexity depends on the received sequence. We compare our decoding algorithm with Chase (1972) algorithm 2 and the Tanaka-Kakigahara (1983) algorithm in which a similar method for generating candidate codewords is used. Computer simulation results indicate, for some signal-to-noise ratios (SNR), that our decoding algorithm requires less average complexity than those of the other two algorithms, but the performance of ours is always superior to those of the other two     

A generalized linear quasi-ML decoder of OSTBCs for wireless communications over time-selective fading channels

We propose a novel generalized linear quasi-maximum-likelihood (quasi-ML) decoder for orthogonal space-time block codes (OSTBCs) for wireless communications over time-selective fading channels. The proposed decoder computes the decision statistics based on the channel-state information and completely removes the intertransmit-antenna interference to provide excellent diversity advantage when the channel varies from symbol to symbol. It is shown that when the channel is quasi-static, the proposed decoder is the optimum ML decoder for OSTBCs. The theoretical bit-error probabilities of the proposed decoder are given and it is shown that the proposed decoder does not exhibit error floors at high signal-to-noise ratios like the decoder proposed in and . Simulation results for various channel-fading rates are presented to verify the theoretical analysis.     

A programmable decoder for Reed-Solomon codes

A programmable decoder for Reed-Solomon codes is described. The decoder is constructed using the Am2900 family of bit-slice elements and it is designed to perform error-correction erasure-filling decoding. The decoding rates obtained, lie in the range of 87 to 22 kilobits per second depending on the code and the error-erasure pattern encountered     

Optimal contention window size for IEEE 802.15.3c mmWave WPANs

The millimeter-wave (mmWave) band offers the potential for multi-gigabit indoor Wireless Personal Area Networks (WPANs). However, it has problems such as short communication coverage due to high propagation losses. In order to compensate for this drawback, utilization of directional antennas at the physical layer is highly recommended. In this paper, we consider the adequate contention window (CW) size for directional carrier sense multiple access with collision avoidance (CSMA/CA). To find the optimal CW size that enhances the performance of conventional directional CSMA/CA, we propose an enhanced directional CSMA/CA algorithm. The algorithm is considered in IEEE 802.15.3c, a standard for mmWave WPANs, under saturation environments. For the algorithm, we present a Markov chain model and analyze it for the no-ACK mode. The effects of directional antennas and the features of IEEE 802.15.3c Medium Access Control (MAC) such as backoff counter freezing are considered in the model. The optimal CW sizes for the two different objective functions are derived from the numerical results. The numerical results also show that the system throughput and average transmission delay of the proposed algorithm outperform those of conventional one and the overall analysis is verified by simulation. The obtained results provide the criterion for selecting the optimal parameters and developing a MAC protocol that enhances the performance of mmWave WPANs.     

纹理分析窗大小的高斯-马尔可夫随机场模型估计方法

在纹理分析中,窗口大小的选择对所提取特征的有效性及计算速度等有很大影响。文中利用高斯－马尔可夫随机场（ＧＭＲＦ）模型对纹理进行描述,采用最小平方误差估计获取纹理图像的随机场参数,并证明了这种估计的一致性。针对估计式在某些情况下可能无解,对该式作了改进,使其在实际应用中总能有解。利用估计的一致性,提出了一种系统估计纹理分析窗口大小的方法,实验表明了这种方法的有效性。     

A simplified fixed-complexity sphere decoder for V-BLAST systems

A simplified fixed-complexity sphere decoder (SFSD) with much lower computational cost than that of the original fixed-complexity sphere decoder (FSD) has been developed for signal detection in vertical Bell Laboratories layered space-time (V-BLAST) systems. Simulation results on a 16-QAM system with 8 transmit and 8 receive antennas show that the SFSD can attain the near-optimal performance with a cost that is about 37% of the counterpart of the original FSD.     

On general linear block code decoding using the sum-product iterative decoder

The sum-product iterative decoder, conventionally used for low-density parity-check (LDPC) codes, hold promise as a decoder for general linear block code decoding. However, the promise is only partly fulfilled because, as we show experimentally, the decoder performance degrades rapidly as a function of parity check matrix weight. Even in the case of decoder failure, however, we demonstrate that there is information present in the decoder output probabilities that can still help with the decoding problem.     

A universal lattice code decoder for fading channels

We present a maximum-likelihood decoding algorithm for an arbitrary lattice code when used over an independent fading channel with perfect channel state information at the receiver. The decoder is based on a bounded distance search among the lattice points falling inside a sphere centered at the received point. By judicious choice of the decoding radius we show that this decoder can be practically used to decode lattice codes of dimension up to 32 in a fading environment     

A low-power dual-mode video decoder for mobile applications

The objective of this article is to highlight design challenges for low-power and dual-video standard requirements, especially in mobile applications. Due to the advent of the newly announced H.264, a generic problem of standard incompatibility has appeared between H.264 and prevalent MPEG-x video standards, which must be resolved on both algorithmic and architectural levels. Furthermore, several low-power techniques targeted at achieving lower memory requirements and processing cycles are also described and discussed.     

A VLSI design for a trace-back Viterbi decoder

A systolic Viterbi decoder for convolutional codes is developed which uses the trace-back method to reduce the amount of data needed to be stored in registers. It is shown that this new algorithm requires a smaller chip size and achieves a faster decoding time than other existing methods     

A high-speed low-complexity Reed-Solomon decoder for optical communications

This paper presents a high-speed low-complexity Reed-Solomon (RS) decoder architecture using a novel pipelined recursive modified Euclidean (PrME) algorithm block for very high-speed optical communications. The RS decoder features a low-complexity key equation solver using a PrME algorithm block. The recursive structure enables the novel low-complexity PrME algorithm block to be implemented. Pipelining and parallelizing allow the inputs to be received at very high fiber-optic rates, and outputs to be delivered at correspondingly high rates with minimum delay. This paper presents the key ideas applied to the design of an 80-Gb/s RS decoder architecture, especially that for achieving high throughput and reducing complexity. The 80-Gb/s 16-channel RS decoder has been designed and implemented using 0.13-/spl mu/m CMOS technology in a supply voltage of 1.2 V. The proposed RS decoder has a core gate count of 393 K and operates at a clock rate of 625 MHz.     

Extension of source coding theorems for block codes to sliding-block codes

For arbitrary alphabets and single-letter fidelity criteria, two theorems are given which allow any fixed-rate or variable-rate source coding theorem for block codes to be extended to sliding-block codes. Applications are given to universal coding and to the coding of a stationary nonergodic source.     

A digital array based bit serial processor for arbitrary window size kernel convolution in vision sensors

The high speed and in-pixel processing of image data in smart vision sensors is an important solution for real time machine vision tasks. Diverse architectures have been presented for array based kernel convolution processing, many of which use analog processing elements to save space. In this paper a digital array based bit serial architecture is presented to perform certain image filtering tasks in the digital domain and hence gain higher accuracies than the analog methods. The presented method benefits from more diverse convolution options such as arbitrary size kernel windows, compared with the digital pulse based approaches. The proposed digital cell structure is compact enough to fit inside an image sensor pixel. When incorporated in a vision chip, resolutions of up to 12 bit accuracy can be obtained in kernel convolution functions with 35×28 μm² layout area usage per pixel in a 90 nm technology. Still, higher accuracies can be obtained with larger pixels. The power consumption of the approach is approximately 10 nW/pixel at a frame rate of 1 kfps.     

A maximum-likelihood bound approaching minimum bit error rate linear multiuser detector for frequency-selective DS-CDMA systems

In this letter, a minimum bit error rate (MBER) linear multiuser detector (MUD) is considered for direct-sequence code-division multiple-access (DS-CDMA) communication systems, distorted by time-varying and frequency-selective multipath fading channels. Based on the approach for finding filter coefficients of the proposed MBER MUD, an efficient Newton method with a barrier parameter is developed. The BER performance of the MBER MUD is compared to other conventional detectors. The study finds that the proposed MBER MUD has more than 2 dB gain over the linear minimum mean-squared error (LMMSE) detector. Furthermore, in the high SNR region, the BER performance of the proposed MBER MUD approaches the performance of the maximum-likelihood (ML) detector.     

Monotonicity of the quantum linear programming bound

The most powerful technique known at present for bounding the size of quantum codes of prescribed minimum distance is the quantum linear programming bound. Unlike the classical linear programming bound, it is not immediately obvious that if the quantum linear programming constraints are satisfiable for dimension K, then the constraints can be satisfied for all lower dimensions. We show that the quantum linear programming bound is monotonic in this sense, and give an explicitly monotonic reformulation     

CCD for PAL decoder

The ultrasonic glass delay line section of PAL color TV receiver is incorporated into an IC by using a CCD delay line. It is shown that it is possible to delay, demodulate, and matrix the chroma signal from the chrominance filter all in the charge domain in one complex CCD which is driven by clock pulses with phases according to the color subcarrier. The output signals are fed directly to the RGB output drivers. Thus the whole chroma section of a PAL-TV set consists only of one MOS-CCD-IC performing the color signal processing and one bipolar IC with the 4.4 MHz oscillator, PAL switch, etc., without external glass delay line and coils.     

A versatile time-domain Reed-Solomon decoder

A versatile Reed-Solomon (RS) decoder structure based on the time-domain decoding algorithm (transform decoding without transforms) is developed. The algorithm is restructured, and a method is given to decode any RS code generated by any generator polynomial. The main advantage of the decoder structure is its versatility, that is, it can be programmed to decode any Reed-Solomon code defined in Galois field (GF) 2^m with a fixed symbol size m. This decoder can correct errors and erasures for any RS code, including shortened and singly extended codes. It is shown that the decoder has a very simple structure and can be used to design high-speed single-chip VLSI decoders. As an example, a gate-array-based programmable RS decoder is implemented on a single chip. This decoder chip can decode any RS code defined in GF (2⁵) with any code word length and any number of information symbols. The decoder chip is fabricated using low-power 1.5-μ, two-layer-metal, HCMOS technology     

一种面向 DVB-C的单片MPEG-2解码方案

介绍了一种面向DVB－C机顶盒的单片MPEG－2解码方案。首先简介了主芯片MB8712250的几个主要部分,然后分别从硬件和软件介绍了方案中各个模块的实现。