Search

Blind source separation (BSS) problem is an open area of research that requires further investigations. Various algorithms were presented in the literature based on second-order statistics and higher-order statistics. The computational complexity of those methods is a challenging task and must be carefully considered to produce fast BSS algorithms. In blind source extraction (BSE) using linear predictors, the adaptive filter update requires complex computations that need consideration. This work focus on new BSE using the idempotent transformation matrix. New algorithm is presented in this work to compute the matrix with less computational complexity as compared with the standard singular value decomposition method. New optimization problem was defined according to the proposed matrix equation, and solved by an iterative algorithm with low computational complexity. The proposed method is tested using speech and white Gaussian signals. The performance measures used in this work are the signal-to-interference ratio, signal-to-distortion ratio, and signal-to-artifact ratio. Simulation results show that the proposed algorithm significantly separate the source signals with better performance measures as compared with the state-of-the-art approaches such as second-order blind identification and fast independent component analysis. [ABSTRACT FROM AUTHOR]

Abstract: An efficient method for face recognition which is robust under illumination variations is proposed. The proposed method achieves the illumination invariants based on the illumination-reflection model employing local matching for best classification. Different filters have been tested to achieve the reflectance part of the image, which is illumination invariant, and maximum filter is suggested as the best method for this purpose. A set of adaptively weighted classifiers vote on different sub-images of each input image and a decision is made based on their votes. Image entropy and mutual information are used as weight factors. The proposed method does not need any prior information about the face shape or illumination and can be applied on each image separately. Unlike most available methods, our method does not need multiple images in training stage to get the illumination invariants. Support vector machines and k-nearest neighbors methods are used as classifier. Several experiments are performed on Yale B, Extended Yale B and CMU-PIE databases. Recognition results show that the proposed method is suitable for efficient face recognition under illumination variations. [Copyright &y& Elsevier]

This paper points out some problems inherent in H.264/AVC intra frame rate control (RC) and presents several new algorithms to solve them. Three main algorithms are proposed in this paper. Multiple quantization parameters (QPs) determination algorithm based on the statistics of the deviation measure is proposed. Thus, relatively accurate QP estimation can be achieved since it takes the features of the current frame content into consideration. Two slice-based RC schemes are also proposed to stabilize the output bit rate. The scene change is detected with the complexity of adjacent frames based on deviation measure, which not only generates stable bit rate output but also improves the video quality. The combination of all the above algorithms improve the average peak signal-to-noise ratio (PSNR) by 0.74 dB while keeping the mismatch of the bit rate output less than 7.18% compared to the H.264/AVC reference model, JM12.0, under various sequences. [ABSTRACT FROM AUTHOR]

Abstract: This paper investigates computationally efficient methods for blind decision feedback equalization (DFE) that reduce the complexity and power requirements of blind equalization algorithms while maintaining their steady-state characteristics for quadrature amplitude modulation (QAM) signals. These include the power-of-two error (POT), selective coefficient update (SCU), and frequency-domain block (FDB) methods. A novel radius-directed stop-and-go (RSG) method is introduced, which selectively adjusts the equalizer tap coefficients based on the equalizer output radius. In addition, a new activation/de-activation method based on the equalizer output radius is utilized to control the feedback equalizer (FBE) of the DFEs. Simulation studies and analysis are provided for empirically derived cable and microwave channels and Ricean fading channels. [Copyright &y& Elsevier]

Lattice recursive least-squares (LRLS) adaptive filters are well known for their fast convergence to achieve the minimum least-squares error. This accelerated decaying error property, however, comes with a high cost filter structure. Using a variable tap-length technique known as the fractional tap-length (FT) algorithm, this article provides an adaptive method to search for the optimum length of LRLS filters. To validate the performance of the proposed algorithm, simulations are conducted under low noise and high noise conditions. All the simulation results show the advantages of the proposed algorithm compared with fixed length LRLS filters as well as with the fractional tap-length lattice based least-mean-squares (FT-LLMS) algorithm. With a fast convergence rate and a small steady-state error of the tap-length, the proposed algorithm can efficiently track any change in the impulse response and estimate the filter's optimum structure. [ABSTRACT FROM AUTHOR]

This paper presents a new non-invasive deterministic algorithm of extracting the fetal Electrocardiogram (FECG) signal based on a new null space idempotent transformation matrix (NSITM). The mixture matrix is used to compute the ITM. Then, the fetal ECG (FECG) and maternal ECG (MECG) signals are extracted from the null space of the ITM. Next, MECG and FECG peaks detection, control logic, and adaptive comb filter are used to remove the unwanted MECG component from the raw FECG signal, thus extracting a clean FECG signal. The visual results from Daisy and Physionet real databases indicate that the proposed algorithm is effective in extracting the FECG signal, which can be compared with principal component analysis (PCA), fast independent component analysis (FastICA), and parallel linear predictor (PLP) filter algorithms. Results from Physionet synthesized ECG data show considerable improvement in extraction performances over other algorithms used in this work, considering different additive signal-to-noise ratio (SNR) increasing from 0 dB to 12 dB, and considering different fetal-to-maternal SNR increasing from −30 dB to 0 dB. The FECG detection of the NSITM is evaluated using statistical measures and results show considerable improvement in the sensitivity (SE), the accuracy (ACC), and the positive predictive value (PPV), as compared with other algorithms. The study demonstrated that the NSITM is a feasible algorithm for FECG extraction. [ABSTRACT FROM AUTHOR]

Bus-Invert ( BI ) coding schemes have been used in the circuit design to save power usage by minimizing the number of transitions between successive words. With an increased number of bits for bus width, many other coding schemes like Bus-Shift ( BS ) and Shift-Invert ( SINV ) have been proposed for better power saving. In this paper, we are proposing a new Successive Shift-Invert ( SSINV ) coding scheme where the transmitting data will first be inverted. Then both the original and the inverted word will be shifted circularly for each bit and compared with the data just sent for least hamming distance. Thus, the combination having the least hamming distance with the last sent data will be selected and sent. The proposed method exhibits better power saving compared to other reported methods, but needs one extra control bit than BS ; which add to the computation time. Simulation results show a saving margin of 5–15% in average cases in comparison with BI , BS , and SINV for higher bus width. A VerilogA implementation of SSINV using TSMC 65 nm technology is also presented as a proof of concept. [ABSTRACT FROM AUTHOR]

This paper presents a new learning-based framework to jointly solve equalization and symbol detection problems in orthogonal frequency division multiplexing systems with quadrature amplitude modulation. The framework utilizes extreme learning machine (ELM), a recent addition to the class of supervised learning algorithms, to achieve fast training, high performance, and low error rates. The proposed ELM scheme employs infinitely differentiable nonlinear activation functions in least-square solution to learn the channel response, which is the equalization part. In addition to equalization, ELM performs symbol detection. Existing learning-based schemes require an additional symbol slicer for the symbol detection. The proposed framework does not experience training bottleneck imposed by gradient descent-based approaches. Simulation results show that the proposed framework outperforms other learning-based equalizers in terms of symbol error rate and training speeds. [ABSTRACT FROM AUTHOR]

Abstract: This paper proposes the use of equalization concepts in frequency domain that exploit the frequency domain channel matrix to combat inter-carrier interference (ICI) instead of inter-symbol interference (ISI) in 5.9GHz Dedicated Short Range Communications (DSRC) systems. The physical layer (PHY) of DSRC is currently being developed by work group of IEEE 802.11p. The conventional system currently assumes static channel characteristics. Channel tracking schemes were investigated and the Viterbi-aided channel estimation scheme was proposed for DSRC systems that did not explicitly exploit the ICI components caused by the time-varying channels. The performance of the DSRC system is investigated for a time-varying channel using a conventional DSRC model, a decision-directed (Viterbi-aided) channel estimation model, and the frequency-domain equalization design. It is shown that the DSRC system with the frequency-domain equalization achieves a considerable performance enhancement compared to both the conventional and the Viterbi-aided channel estimation schemes in terms of both packet error rate (PER) and bit error rate (BER) at relatively high and low velocities. [Copyright &y& Elsevier]

Abstract: In this paper, a design of a new object-oriented simulation environment and graphical user interface (GUI) for versatile orthogonal frequency division multiplexing (OFDM) systems is presented. This is achieved using high-level design, parallelism and usability for the simulation environment. GUI interface can serve as a learning/research tool for students or practiced professionals to investigate particular designs. It is evident that OFDM systems intended to be used in dynamic environments must be tested under various conditions in order to be designed for optimality. Hence, a simulation design is proposed coupling the GUI, parallelism, and high-level object-oriented design techniques to be beneficial to the researcher. A high-level design and GUI layouts of the proposed simulator is shown in details. Important OFDM parameters needed for reconfiguration of transmitter components, channel condition parameters, and receiver components are discussed. In addition, this paper provides a simple technique to implement simulation partitioning for increased parallel performance of reconfigurable object-oriented OFDM simulators. This simple technique applies to scenarios where there is disproportionate simulation duration between different OFDM configurations. It is shown to decrease total simulation time considerably. [Copyright &y& Elsevier]

Abstract: In this paper, we present a study of the dedicated short range communications (DSRC) receiver performance under varying signal to noise ratio, velocity, symbol durations and packet lengths. Conventional channel estimation, which is used in IEEE 802.11a, assumes static channel characteristics for the entire packet duration. That is found to be infeasible for high velocity DSRC applications. Simulation results show that the packet-error-rate increases with the increase in relative velocity. Viterbi decoding substantially improves the performance, but the sensitivity to Doppler shift still exists. Analysis and simulation results show that extending the symbol duration or increasing packet length results in an increase of the packet-error-rate. These results may serve as benchmarks for future DSRC channel estimation methods. [Copyright &y& Elsevier]