No abstract

A monolithic multicomponent system is proposed and implemented on a III-nitride-on-silicon platform, whereby two multiple-quantum-well diodes (MQW-diodes) are interconnected by a suspended waveguide. Both MQW-diodes have an identical low-In-content InGaN/Al0.10Ga0.90N MQW structure and are produced by the same fabrication process flow. When appropriately biased, both MQW-diodes operate under a simultaneous emission-detection mode and function as a transmitter and a receiver at the same time, forming an in-plane full-duplex light communication system. Real-time full-duplex audio communication is experimentally demonstrated using the monolithic multicomponent system in combination with an external circuit.

Neural source-filter (NSF) models are deep neural networks that produce waveforms given input acoustic features. They use dilated-convolution-based neural filter modules to filter sinebased excitation for waveform generation, which is different from WaveNet and flow-based models. One of the NSF models, called harmonic-plus-noise NSF (h-NSF) model, uses separate pairs of source and neural filters to generate harmonic and noise waveform components. It is close to WaveNet in terms of speech quality while being superior in generation speed.The h-NSF model can be improved even further. While h-NSF merges the harmonic and noise components using predefined digital low-and high-pass filters, it is well known that the maximum voice frequency (MVF) that separates the periodic and aperiodic spectral bands are time-variant. Therefore, we propose a new h-NSF model with time-variant and trainable MVF. We parameterize the digital low-and highpass filters as windowed-sinc filters and predict their cut-off frequency (i.e., MVF) from the input acoustic features. Our experiments demonstrated that the new model can predict a good trajectory of the MVF and produce high-quality speech for a text-to-speech synthesis system.

No abstract

No abstract

A novel method employing two-dimensional (2D) unitary ESPRIT and a spatial smoothing technique is proposed for super-resolution ISAR imaging. The spatial smoothing technique is employed to resolve the contradiction that 2D unitary ESPRIT requests multiple snapshots; however, in radar imaging applications only one snapshot of the radar data are available. Experimental results are presented to verify the high resolution and feasibility of the proposed imaging algorithm in ISAR imaging.Introduction: In ISAR imagery applications, conventional range Doppler imaging is implemented by two-dimensional (2D) fast Fourier transform (FFT) with the resolution limited by the bandwidth and the rotation angle during the observation time. Recently, modern spectral estimation technologies have become a kind of new methods. 2D multiple signal classification (2D MUSIC) [1, 2] and 2D ESPRIT [3][4][5] are the most popular super-resolution algorithms for 2D parameter estimation without the pairing step. 2D ESPRIT needs no searching process as does the 2D MUSIC, which reduces the huge computational burden. However, 2D ESPRIT has its own drawbacks such as the fact that the conjugation of observed data is not used and formulated in terms of complex-valued data. In this Letter, we present a 2D unitary ESPRIT-based method to obtain super-resolution ISAR imaging and take full advantage of the observed data.

No abstract

In the magnetic impedance approach for damage detection, there is no direct contact between the sensor and the host structure, and the impedance sensor can be moveable above the structure surface. This has promising aspects, especially for online health monitoring of structures with complicated geometries and boundaries. In an earlier study, we have demonstrated that integrating a synthetic, tunable negative resistance and a tunable capacitor with the magnetic transducer can amplify both the measurement magnitude and the damage-induced impedance anomaly and yield much higher signal-to-noise ratio and detection sensitivity. In this research, we formulate detailed modeling and carry out comprehensive analysis to investigate the enhanced magnetic impedance sensing with circuitry integration. Specifically, aiming at providing design and implementation guidelines, we investigate systematically the magnetomechanical coupling effect in such a sensing scheme, which focuses on elucidating quantitatively the influence of various sensor parameters to the impedance measurements. The modeling and analysis are validated by experimental studies.