Zhigang Cao

Improved MFCC-based feature for robust speaker identification

Abstract The Mel-frequency cepstral coefficient (MFCC) is the most widely used feature in speech and speaker recognition. However, MFCC is very sensitive to noise interference, which tends to drastically degrade the performance of recognition systems because of the mismatches between training and testing. In this paper, the logarithmic transformation in the standard MFCC analysis is replaced by a combined function to improve the noisy sensitivity. The proposed feature extraction process is also combined with speech enhancement methods, such as spectral subtraction and median-filter to further suppress the noise. Experiments show that the proposed robust MFCC-based feature significantly reduces the recognition error rate over a wide signal-to-noise ratio range.

Dynamic spatial subchannel allocation with adaptive beamforming for MIMO/OFDM systems

Orthogonal frequency division multiplexing (OFDM) has been widely regarded as an effective modulation technique for mitigating the effects of intersymbol interference in a frequency selective fading channel and for providing reliable high-data-rate transmission over wireless links. Adaptive antenna arrays at the base and mobile stations can achieve further increases in systems capacity and bandwidth efficiency, as well as in quality-of-service improvement in conventional OFDM systems. The conventional adaptive antenna-arrays-based OFDM systems always use the subcarriers characterized by the largest eigenvalues to transmit the OFDM block symbols. And in contrast to previous work, we propose dynamic spatial subchannel allocation with adaptive beamforming for broadband OFDM wireless transmission systems. The proposed system adaptively selects the eigenvectors associated with the relatively large spatial subchannel eigenvalues to generate the beamforming weights at the mobile and base stations and then dynamically assigns the corresponding best spatial subchannels to transmit the OFDM block symbols. It is shown that the proposed system can achieve better performance than an adaptive antenna-arrays-based OFDM system without dynamic spatial subchannel allocation over multipath fading channels. Simulation results also reveal that the proposed system is far less susceptible to feedback delay in rapid time-varying channels and a little more sensitive to channel estimation errors than conventional adaptive antenna-arrays-based OFDM systems. The performance of the proposed system combined with adaptive modulation is also considered.

Interference cancellation for OFDM systems in presence of overlapped narrow band transmission system

Narrowband interference (NBI) is one of the most important problems in an OFDM system, which not only degrades the signal performance on the overlapped subcarriers, but also affects the signal performance on the nearby subchannels due to the spectral leakage effect of DFT demodulation at the receiver. In this paper, we consider the NBI problem caused by narrow band digital communication systems, and propose a novel NBI cancellation method. We consider the NBI reconstruction problem including NBI central frequency and its bandwidth estimation. Simulation results show that this method can achieve about 6 dB signal to interference plus noise ratio (SINR) gain if the interference bandwidth is about twice of that OFDM bins and the system SINR is about 0 dB. As the bandwidth of NBI decreases, more SINR gain will be obtained. Meanwhile, we also consider the system implementation problem, it has shown that its implementation complexity is not high based on the current digital signal processing techniques. If zoom FFT is employed in our system, the calculation complexity can he reduced considerably.

Dual-Microphone Source Location Method in 2-D Space

Array based source location attracts a growing interest nowadays, which is frequently used in videoconference, hearing aids and hands free telephone systems to detect a speakers position. Time delay estimation (TDE) based dual-step source location is assumed to be classical method in this field. Recently, a new technique is developed based on interaural level difference (ILD) method, which determines the source position by the energy ratio from microphone pairs. However, all these localization techniques need at least three sensors to obtain a 2D source locus. In this work, we combine the former two techniques, i.e., ILD and TDE based techniques, to present a novel localization approach by an array of only two microphones, and further provide its closed form solution. Our final simulation confirms that such method, which is thought to be more suitable for the equipments with small size, can achieve a good result under normal conditions

A novel narrowband interference canceller for OFDM systems

Narrowband interference (NBI) will degrade the performance in an OFDM system not only on the overlapped subcarriers, but also on the nearby subchannels due to the spectral leakage effect of DFT demodulation. In this paper we proposed a novel NBI suppression method in the case that NBI is caused by a narrowband digital communication system. We estimate the transmitted data of NBI signal and reconstruct its waveform, by measuring interference information on certain unmodulated subcarriers. And then subtract estimated disturbance in frequency domain. Simulation results show that this method can achieve an average SINR gain about 6 dB on a multipath fading channel, when the interference has equal power with the desired signal and twice bandwidth of an OFDM bin. With less bandwidth of NBI, more performance gain will be obtained.

A neighbor-table-based multipath routing in ad hoc networks

Due to station mobility, the network topology is continuously changing in mobile ad hoc networks. In this paper, we propose a neighbor-table-based multipath routing (NTBMR) protocol to track the dynamic topology changes. Distinguished from prior work on multipath protocols employing disjoint paths, NTBMR does not require the routes to be disjoint. In order to verify the different capabilities against dynamic topology changes for disjoint and non-disjoint multipath routings, we make an attempt to analyze their route reliabilities. Theoretical analysis reveals that non-disjoint multipath routing has higher route reliability when the wireless links are unreliable. In NTBMR scheme, we also present a technique to estimate the mean and variance of the lifetime of a wireless link, which can be used to aid route discovery and maintenance. Simulation results show that our multipath routing scheme is relatively robust in an environment with frequent topology changes and can improve the end-to-end delay and packet delivery ratio performance substantially compared to unipath routing.

Dynamic sub-channel allocation with adaptive beamforming for broadband OFDM wireless systems

Orthogonal frequency division multiplexing (OFDM) has been widely regarded as an effective modulation technique for mitigating the effects of ISI in a frequency selective fading channel and for providing reliable high data rate transmission over wireless links. Adaptive antenna arrays at the base and mobile stations can achieve further increases in systems capacity and bandwidth efficiency, as well as in QoS improvement in conventional OFDM systems. The conventional adaptive antenna arrays based OFDM systems always use the sub-carriers characterized by the first largest eigenvalues to transmit the OFDM block symbols. In this paper and in contrast to previous work, we propose dynamic sub-channel allocation with adaptive beamforming for broadband OFDM wireless transmission systems. The proposed system adaptively selects the eigenvectors associated with the relatively large sub-channel eigenvalues to generate the beamforming weights at the mobile and base stations and then dynamically assigns the corresponding best sub-channels to transmit the OFDM block symbols. It is shown by simulation that, without adding much complexity, the proposed system can achieve a better performance than an adaptive antenna arrays based OFDM system without dynamic sub-channel allocation over multipath fading channels. Simulation results also reveal that the proposed system is not too sensitive to channel estimation errors.

Receiver window design for narrowband interference suppression in IEEE 802.11a system

Orthogonal frequency division multiplexing (OFDM) has been adopted as the physical layer transmission technology in IEEE 802.11a standard, where the Fourier transform is used in demodulation. However, because of the effect of spectral leakage in discrete Fourier transform, narrowband interference is spread in the spectrum, thus, strongly impacts adjacent subchannels. Receive window function was proposed previously to suppress narrow band interference spreading in OFDM systems. We take the symmetry of the long preamble in IEEE 802.11a physical layer frame into account and propose a Hanning window based channel estimation. We also investigate the selection of window function for data reception based on the measurements of indoor wireless environment The joint windows-based receive scheme satisfies the 802.11a standard and requires no modification at the transmitter. Theoretical analysis and simulations demonstrate the effectiveness of this technique in improving the system performance.

Adaptive beamforming with antenna selection in MIMO systems

Selection diversity at both the transmitter and receiver is the simplest way to realize MIMO (multiple-input multiple-output) systems, but the performance improvement of this form of diversity technique is limited. Adaptive beamforming based MIMO can provide significant increases in a systems capacity and performance, but it is characterized by a relatively higher implementation complexity than selection diversity. We propose an adaptive beamforming with antenna selection for MIMO wireless transmission systems. By adaptively selecting the best beamforming which corresponds to the largest eigenvalue in the subset of all available channel correlation matrices, the proposed system can get significant performance improvements over the conventional adaptive antenna array based MIMO systems and the selective space-time block coding approach over flat fading channels. Numerical results also show that the proposed approach always outperforms the selective space-time coding approach under imperfect CSI.

An enhanced CSMA-CA mechanism for multihop ad hoc networks

We propose a two-dimensional (2D) Markov model, which considers packet losses, for carrier sense multiple access with collision avoidance (CSMA-CA) mechanism. Since the initial contention window (ICW) has great impact on the performance of CSMA-CA, we introduce an adjustable state parameter q(0<q<1) to extend ICW. In addition, we analyze the number of competing stations for any source station in a multihop ad hoc network. Based on the analysis result, we develop an adjusting procedure for CSMA-CA mechanism in multihop ad hoc networks. The simulation results show that the packet loss ratio can be reduced through dynamically adjusting ICW by changing q.