Borching Su

National Taiwan University | NTU · Graduate Institute of Communication Engineering

Beamforming techniques are popular in speech-related applications due to their effective spatial filtering capabilities. Nonetheless, conventional beamforming techniques generally depend heavily on either the target's direction-of-arrival (DOA), relative transfer function (RTF) or covariance matrix. This paper presents a new approach, the intelligi...

Speech enhancement (SE) approaches can be classified into supervised and unsupervised categories. For unsupervised SE, a well-known cycle-consistent generative adversarial network (CycleGAN) model, which comprises two generators and two discriminators, has been shown to provide a powerful nonlinear mapping ability and thus achieve a promising noise...

Sparse code multiple access (SCMA), as a codebook-based non-orthogonal multiple access (NOMA) technique, has received research attention in recent years. The codebook design problem for SCMA has also been studied to some extent since codebook choices are highly related to the system's error rate performance. In this paper, we approach the downlink...

For speech-related applications in Internet of things environments, identifying effective methods to handle interference noises and compress the amount of data in transmissions is essential for achieving high-quality services. In this paper, we propose a novel multi-input multi-output speech compression and enhancement (MIMO-SCE) system based on a...

In this article, a scalable hybrid phased-array system is presented through synchronization, analog complex weighting, and digital beamforming of numerous fully integrated Ka-band four-receiver (RX)/four-transmitter (TX) phased-array transceiver integrated circuits (ICs). A 1.09-GHz clock synchronizes the local oscillator (LO) and a 50-MHz clock sy...

Generalized frequency division multiplexing (GFDM) is a promising candidate waveform for next-generation wireless communication systems. However, GFDM channel estimation is still challenging due to the inherent interference. In this paper, we formulate a pilot design framework with linear minimum mean square error (LMMSE) channel estimation for GFD...

Abstract One of the demands of the next-generation wireless communication systems is being supportive to asynchronous traffic types. In order to meet this demand, many waveform candidates for next-generation wireless communication systems have low out-of-subband emissions (OOSBE) of the transmitted signal, and the popular ones deploy either filteri...

A downlink (DL) precoding scheme that needs no DL training nor uplink (UL) feedback for massive MIMO systems operated under frequency-division duplex (FDD) is proposed, whereas most of the schemes in the literature were proposed to compress the overhead of training and feedback. Due to the above advantage, the proposed scheme is suitable for extrem...

A new circularly pulse-shaped (CPS) precoding orthogonal frequency division multiplexing (OFDM) waveform, or CPS-OFDM for short, is proposed in this paper. CPS-OFDM, characterized by user-specific precoder flexibility, possesses the advantages of both low out-of-subband emission (OSBE) and low peak-to-average power ratio (PAPR), which are two major...

Circularly pulse-shaped orthogonal frequency division multiplexing (CPS-OFDM) is one of the most promising 5G waveforms that addresses two physical layer signal requirements of low out-of-subband emission (OSBE) and low peak-to-average power ratio (PAPR) with flexibility in parameter adaptation. In this paper, a constellation shaping optimization m...

A new circularly pulse-shaped (CPS) precoding orthogonal frequency division multiplexing (OFDM) waveform, or CPS-OFDM for short, is proposed in this paper. CPS-OFDM, characterized by user-specific precoder flexibility, possesses the advantages of both low out-of-subband emission (OSBE) and low peak-to-average power ratio (PAPR), which are two major...

Generalized frequency division multiplexing (GFDM) is considered a non-orthogonal waveform and known to encounter difficulties when using in the spatial multiplexing mode of multiple-input-multiple-output (MIMO) scenario. In this paper, a class of GFDM prototype filters, under which the GFDM system is free from inter-subcarrier interference, is inv...

Distributed speech recognition (DSR) splits the processing of data between a mobile device and a network server. In the front-end, features are extracted and compressed to transmit over a wireless channel to a back-end server, where the incoming stream is received and reconstructed for recognition tasks. In this paper, we propose a feature compress...

Different from conventional mobile networks designed to optimize the transmission efficiency of one particular service (e.g., streaming voice/ video) primarily, the industry and academia are reaching an agreement that 5G mobile networks are projected to sustain manifold wireless requirements, including higher mobility, higher data rates, and lower...

Recently, a subspace-constrained diagonal loading (SSC-DL) method has been proposed for robust beamforming against the mismatched direction of arrival (DoA) issue. Although SSC-DL has outstanding output SINR performance, it is not clear how to choose the DL factor and subspace dimension in practice. The aim of the present study is to further invest...

Generalized frequency division multiplexing (GFDM) is considered to be a promising waveform candidate for 5G new radio. It features good properties such as low out-of-band (OOB) radiation. One major drawback of GFDM known in the literature is that a zero-forcing receiver suffers from the noise enhancement effect since the GFDM (transmitter) matrix...

This paper presents a fully-integrated Ka-band 4RX/4TX phased-array transceiver IC. The four transceiver channels are synchronized by an I/Q standing wave oscillator (SWO). Their signal phases are shifted by hybrid analog IF and LO phase interpolators, and can be controlled digitally and independently. The SWO further composes an on-chip PLL, and i...

Massive MIMO can provide downlink access to multiple user equipments (UEs) through appropriate precoding or beamforming. To obtain precoding matrices for users, channel state information at the transmitter (CSI-T) is usually mandatory, requiring downlink training and CSI feedback at least in the frequency division duplex mode. However, such trainin...

In this study, a new subspace-constrained diagonal loading (SSC-DL) method is presented for robust beamforming against the issue of a mismatched direction of arrival (DoA), based on an extension to the well known diagonal loading (DL) technique. One important difference of the proposed SSC-DL from conventional DL is that it imposes an additional co...

For most of the state-of-the-art speech enhancement techniques, a spectrogram is usually preferred than the respective time-domain raw data since it reveals more compact presentation together with conspicuous temporal information over a long time span. However, the short-time Fourier transform (STFT) that creates the spectrogram in general distorts...

This paper investigates the use of the speech parameter generation (SPG) algorithm, which has been successfully adopted in deep neural network (DNN)-based voice conversion (VC) and speech synthesis (SS), for incorporating temporal information to improve the deep denoising auto-encoder (DDAE)-based speech enhancement. In our previous studies, we hav...

Orthogonal frequency division multiplexing (OFDM) with offset quadrature amplitude modulation (OQAM) has been considered as a potential alternative to the popular OFDM systems using cyclic prefixes (CP) for advantages in bandwidth efficiency. However, most current channel estimation methods for OFDM/OQAM require a considerable amount of pilots or p...

In this paper, we studied the problem of DOA estimation in the presence of mutual coupling. The proposed method for reconstructing the steering vector can estimate the DOAs in the situation of a large number of source signals with unknown mutual coupling. The main advantage of the proposed method is that the capacity of the number of source signals...

In order to enable frequency-division duplex (FDD) massive multiple-input-multiple-output (MIMO) systems, a beam-grouping, precoding and feedback scheme is proposed to reduce the downlink training overhead and channel state information (CSI) feedback overhead. When the number of transmit antennas on a base station (BS) is growing large, the downlin...

In this paper, a Cramér-Rao bound (CRB) for blind channel estimation in cyclic prefixed MIMO-OFDM systems is derived. The derived bound is valid for arbitrary numbers of transmit antennas, received antennas, and received symbols available for blind channel estimation, and serve as a benchmark of performances for several recently reported subspace-b...

In this paper, we investigate channel estimation methods in carrier aggregation (CA) orthogonal frequency division multiplexing (OFDM) systems. Carrier aggregation (CA) is one of the most distinct features of 4G that allows operators to deploy a system with larger bandwidth by aggregating several smaller contiguous or non-contiguous carriers. By ta...

This study proposes a polynomial based feature transferring (PFT) algorithm for acoustic feature conversion. The PFT process consists of estimation and conversion phases. The estimation phase aims to compute a polynomial based transfer function using only a small set of parallel source and target features. With the estimated transfer function, the...

In this paper, we study subspace-based (SS) methods for blind and semiblind channel estimation in OFDM systems with virtual carriers (VC) using few received received symbols. The methods work in both zero-padded (ZP) and cyclic-prefixed (CP) guard interval types, based on viewing the employment of VCs as precoding of the data matrix using a rank-de...

In this paper, two subspace (SS) based blind channel estimation methods for cyclic prefixing (CP) based multi-input and multi-output (MIMO) OFDM systems are proposed. The first one generalized a previously reported single-input and single-output (SISO) based method into the MIMO case. The second one is an extension to the first one by smartly intro...

A Cramer-Rao bound (CRB) for semi-blind channel estimators in redundant block transmission systems is derived. The derived CRB is valid for any system adopting a full-rank linear redundant precoder, including the popular cyclic-prefixed orthogonal frequency-division multiplexing system. Simple forms of CRBs for multiple complex parameters, either u...

Blind channel estimation has been studied in various types of communication systems since it has a better bandwidth efficiency than training-based counterpart. In communication systems employed with cyclic prefixes (CP), subspace-based (SS) methods are among the most popular categories of blind channel estimation schemes. Existing SS methods, howev...

In this paper, we derive the Cramer-Rao bound (CRB) for blind channel estimation in redundant block transmission systems, a lower bound for the mean squared error of any blind channel estimators. The derived CRB is valid for any full-rank linear redundant precoder, including both zero-padded (ZP) and cyclic-prefixed (CP) precoders. A simple form of...

This paper studies the blind block synchronization problem in block transmission systems using linear redundant precoders (LRP). Two commonly used LRP systems, namely, zero padding (ZP) and cyclic prefix (CP) systems, are considered in this paper. In particular, the block synchronization problem in CP systems is a broader version of timing synchron...

In orthogonal frequency division multiplexing (OFDM) systems, symbol synchronization is a critical step for successful data transmission. While this task is done in most current systems by using training symbols, a few studies have been dedicated to solving the problem blindly, that is, where training symbols are not available. Blind symbol synchro...

In the above-mentioned paper, a Cramer-Rao bound was derived for the performance of a blind channel estimation algorithm. In this correspondence, an error in the bound is pointed out and corrected. It is observed here that the performance of the said algorithm does not achieve the Cramer-Rao bound.

Digital signal processing (DSP) techniques have played an important role in channel equalization and estimation in communication systems. While channel equalization and estimation are usually done by pilot-assisted methods in most systems, algorithms for blind channel estimation have also been largely studied due to high bandwidth efficiency. Howev...

In this paper we study a new semi-blind channel estimation algorithm in orthogonal frequency division multiplexing (OFDM) systems. The proposed scheme is an extension of a recently reported subspace-based blind channel estimation algorithm in cyclic prefix systems which requires very few received blocks. The semi-blind estimation algorithm is devis...

In this letter, we analyze the performance of a recently reported generalized blind channel estimation algorithm. The algorithm has a parameter called repetition index, and it reduces to two previously reported special cases when the repetition index is chosen as unity and as the size of received blocks, respectively. The theoretical performance of...

In this paper, a novel generalization of subspace-based blind channel identification methods in cyclic prefix (CP) systems is proposed. For the generalization, a new system parameter called repetition index is introduced whose value is unity for previously reported special cases. By choosing a repetition index larger than unity, the number of recei...

Recently, a new subspace-based blind channel estimation algorithm in cyclic prefix (CP) system was reported. A persistency of excitation (PE) property of the input signal is required for the algorithm to work. In this paper, the probability of fulfilling the PE property under different situations is studied. Four factors in the algorithm affect the...

Abstract, It is well-known that redundant filter bank pre-coders can be used for blind identification as well as equalization of FIR channels. Several algorithms have been proposed in the literature exploiting trailing zeros in the transmitter. In this paper we propose a generalized algorithm of which the previous algorithms are special cases. By c...

In this paper, a theoretical problem arising in digital communications, namely the generalized signal richness preservation problem, is addressed and studied. In order to solve the problem, a special class of square matrices, namely the "Vandermonde-form preserving" (VFP) matrices, is introduced and found to be highly relevant to the problem. Sever...

Blind channel identification using linear redundant filterbank precoders (LRP) has been studied extensively in the literature. Most methods are proposed based on the assumption that block synchronization is perfect. In practice, a blind block synchronization algorithm must be used to justify this assumption. This paper studies the blind block synch...

In this paper, a theoretical problem arising in digital communications, namely the generalized signal richness preservation problem, is addressed and studied. In order to solve the problem, a special class of square matrices, namely the "Vandermonde-form preserving" (VFP) matrices, is introduced and found to be highly relevant to the problem. Sever...

In this paper a novel generalization of subspace- based blind channel identification methods in cyclic prefix (CP) systems is proposed. For the generalization, a new system parameter called repetition index is introduced whose value is unity for previously reported special cases. By choosing a repetition index larger than unity, the number of recei...

It is well-known that redundant filter bank precoders can be used for blind identification as well as equalization of FIR channels. Several algorithms have been proposed in the literature exploiting trailing zeros in the transmitter. In this paper we propose a generalized algorithm of which the previous algorithms are special cases. By carefully ch...

In this paper, theoretical issues about linear time invariant (LTI) systems that preserve signal richness are explored. This paper considers two particular definitions of signal richness and finds the necessary and sufficient conditions under which an LTI system preserves the richness property. Several examples are presented to clarify the issues i...

The idea of signal richness has recently been studied in considerable details, and the conditions on linear time invariant (LTI) systems to preserve such richness have been established. In this paper, the concept of richness is extended to practically useful classes such as bounded signals. The class of LTI systems studied for the richness preserva...

Antipodal paraunitary (APU) matrices have been shown to be useful with respect to bit error rate (BER) as linear precoders for OFDM systems, especially in a fast-varying channel environment. In this paper we study a broader class of paraunitary matrices, namely, the distributed antipodal paraunitary (DAPU) matrices, of which APU matrices are specia...

There are many ways to define richness of a discrete time signal. In this paper, we consider a particular definition and explore the conditions under which a linear time invariant (LTI) system preserves the richness property. A set of necessary and sufficient conditions has been found. Using this, paraunitary and unimodular matrices can be shown no...

There are many ways to define richness of a discrete time signal. This paper considers a particular definition and explores the conditions under which a linear time invariant system preserves the richness property. Several examples are presented to clarify the issues involved in the problem. Some sufficient conditions are presented. Also presented...

This paper discusses a number of issues pertaining to blind identification of channels. The basics of blind identification are first discussed and a method called Vandermonde method is presented which is based on elementary linear system principles. Then some remarks are made about precoders with paraunitary antipodal preprocessors. It is argued th...

This paper presents a new method for musical chord recognition based on a model of human perception. We classify the chords directly from the sound without the information of timbres and notes. A wavelet-based transform as well as a self-organized map (SOM) neural network is adopted to imitate human ears and cerebra, respectively. The resultant sys...

Blindchannel identification usinglinear redundant filterbank precoders (LRP)hasbeenstudied extensively inthe literature. Mostmethodsareproposed basedontheassumption thatblocksynchronization isperfect. Inpractice, ablind block synchronization algorithm mustbeusedtojustify thisassump- tion. Thispaperstudies theblind blocksynchronization problem insys...