BioCPPNet: automatic bioacoustic source separation with deep neural networks

We introduce the Bioacoustic Cocktail Party Problem Network (BioCPPNet), a lightweight, modular, and robust U-Net-based machine learning architecture optimized for bioacoustic source separation across diverse biological taxa. Employing learnable or handcrafted encoders, BioCPPNet operates directly on the raw acoustic mixture waveform containing overlapping vocalizations and separates the input waveform into estimates corresponding to the sources in the mixture. Predictions are compared to the reference ground truth waveforms by searching over the space of (output, target) source order permutations, and we train using an objective function motivated by perceptual audio quality. We apply BioCPPNet to several species with unique vocal behavior, including macaques, bottlenose dolphins, and Egyptian fruit bats, and we evaluate reconstruction quality of separated waveforms using the scale-invariant signal-to-distortion ratio (SI-SDR) and downstream identity classification accuracy. We consider mixtures with two or three concurrent conspecific vocalizers, and we examine separation performance in open and closed speaker scenarios. To our knowledge, this paper redefines the state-of-the-art in end-to-end single-channel bioacoustic source separation in a permutation-invariant regime across a heterogeneous set of non-human species. This study serves as a major step toward the deployment of bioacoustic source separation systems for processing substantial volumes of previously unusable data containing overlapping bioacoustic signals.

BioCPPNet: Automatic Bioacoustic Source Separation with Deep Neural Networks

We introduce the Bioacoustic Cocktail Party Problem Network (BioCPPNet), a lightweight, modular, and robust UNet-based machine learning architecture optimized for bioacoustic source separation across diverse biological taxa. Employing learnable or handcrafted encoders, BioCPPNet operates directly on the raw acoustic mixture waveform containing overlapping vocalizations and separates the input waveform into estimates corresponding to the sources in the mixture. Predictions are compared to the reference ground truth waveforms by searching over the space of (output, target) source order permutations, and we train using an objective function motivated by perceptual audio quality. We apply BioCPPNet to several species with unique vocal behavior, including macaques, bottlenose dolphins, and Egyptian fruit bats, and we evaluate reconstruction quality of separated waveforms using the scale-invariant signal-to-distortion ratio (SI-SDR) and downstream identity classification accuracy. We consider mixtures with two or three concurrent conspecific vocalizers, and we examine separation performance in open and closed speaker scenarios. To our knowledge, this paper redefines the state-of-the-art in end-to-end single-channel bioacoustic source separation in a permutation-invariant regime across a heterogeneous set of non-human species. This study serves as a major step toward the deployment of bioacoustic source separation systems for processing substantial volumes of previously unusable data containing overlapping bioacoustic signals.

A Novel Wearable Flexible Dry Electrode Based on Cowhide for ECG Measurement

The electrocardiogram (ECG) electrode, as a sensor, is an important part of the wearable ECG monitoring device. Natural leather is rarely used as the electrode substrate. In this paper, wearable flexible silver electrodes based on cowhide were prepared by sputtering and brush-painting. A signal generator, oscilloscope, impedance test instrument, and ECG monitor were used to build the test platform evaluating the performance of electrodes with six subjects. The lossless waveform transmission can be achieved with our electrodes. Therefore, the Pearson’s correlation coefficient calculated with input waveform and output waveform of the electrodes based on the top grain layer (GLE) and the split layer (SLE) of cowhide were 0.997 and 0.998 at 0.1 Hz respectively. The skin electrode impedance (Z) was tested, and the parameters of the equivalent circuit model of the skin electrode interface were calculated by a fitting method, indicating that the Z of the prepared electrodes was comparable with the standard gel electrode when the skin is moist enough. The signal-to-noise ratio of the ECG of the GLE and the SLE were 1.148 and 1.205 times that of the standard electrode in the standing posture, which meant the ECG measured by our electrodes was basically consistent with that measured by the standard electrode.

Unjuk Kerja Filter Pasif dalam Mereduksi Distorsi Gelombang pada AC/DC Konverter

This paper shows the use of passive filters to reduce the distortion of input waves in a single phase uncontrolled rectifier. Simulations are carried out with various passive filter models to get a decrease in the value of Total Harmonic Distortion (THD) and correct the distortion of the uncontrolled single phase rectifier input waveform. The results of the simulation show the shape of the input signal changes in the uncontrolled single-phase rectifier, analyzing the value of power factor improvement and also the THD value.

Beyond K-complex binary scoring during sleep: probabilistic classification using deep learning

Study Objectives K-complexes (KCs) are a recognized electroencephalography marker of sensory processing and a defining feature of sleep stage 2. KC frequency and morphology may also be reflective of sleep quality, aging, and a range of sleep and sensory processing deficits. However, manual scoring of K-complexes is impractical, time-consuming, and thus costly and currently not well-standardized. Although automated KC detection methods have been developed, performance and uptake remain limited. Methods The proposed algorithm is based on a deep neural network and Gaussian process, which gives the input waveform a probability of being a KC ranging from 0% to 100%. The algorithm was trained on half a million synthetic KCs derived from manually scored sleep stage 2 KCs from the Montreal Archive of Sleep Study containing 19 healthy young participants. Algorithm performance was subsequently assessed on 700 independent recordings from the Cleveland Family Study using sleep stages 2 and 3 data. Results The developed algorithm showed an F1 score (a measure of binary classification accuracy) of 0.78 and thus outperforms currently available KC scoring algorithms with F1 = 0.2–0.6. The probabilistic approach also captured expected variability in KC shape and amplitude within individuals and across age groups. Conclusions An automated probabilistic KC classification is well suited and effective for systematic KC detection for a more in-depth exploration of potential relationships between KCs during sleep and clinical outcomes such as health impacts and daytime symptomatology.

Integrated Dwt-Differentiation Algorithm for Fault Detection and Relay Coordination in Micro Grid

The Bidirectional flow of current makes it difficult to detect fault in the microgrid. The level of fault current changes continuously with change in load, it leads to selectivity and sensitivity issue of relay. In this paper integrated DWT-differentiation algorithm is proposed for fault detection and relay coordination, the input waveform of fault current is proceed with discrete wavelet transform. Time scale function of DWT used to extract exact feature from signal which helps in further effective analysis. The Optimization function of relay is mainly depends on PSM (plug setting multiplier) and TDS (Time dial span). The Fault current used to calculate this parameter are already analyzed from DWT. Standard 9 bus IEEE system is used as reference. Fault is detected at 21 different locations; initially primary protection is activated and secondary protection operates only if first selected pair of relay fails to operate .The differential algorithm select best pair of backup relay and relay coordination is carried out resulting in reduction of operating Time

A Comparison of Filtering Approaches Using Low-Speed DACs for Hardware-in-the-Loop Implemented in FPGAs

The use of Hardware-in-the-Loop (HIL) systems implemented in Field Programmable Gate Arrays (FPGAs) is constantly increasing because of its advantages compared to traditional simulation techniques. This increase in usage has caused new challenges related to the improvement of their performance and features like the number of output channels, while the price of HIL systems is diminishing. At present, the use of low-speed Digital-to-Analog Converters (DACs) is starting to be a commercial possibility because of two reasons. One is their lower price and the other is their lower pin count, which determines the number and price of the FPGAs that are necessary to handle those DACs. This paper compares four filtering approaches for providing suitable data to low-speed DACs, which help to filter high-speed input signals, discarding the need of using expensive high-speed DACS, and therefore decreasing the total cost of HIL implementations. Results show that the selection of the appropriate filter should be based on the type of the input waveform and the relative importance of the dynamics versus the area.

Active Control of Bluff-Body Flows Using Plasma Actuators

Actuators play an important role in modern active flow control technology. Dielectric barrier discharge plasma can be used to induce localized velocity perturbations in air, so as to accomplish modifications to the global flow field. This paper presents a selective review of applications from the published literature with emphasis on interactions between plasma-induced perturbations and original unsteady fields of bluff-body flows. First, dielectric barrier discharge (DBD)-plasma actuator characteristics, and the local disturbance fields these actuators induce into the exterior flow, are described. Then, instabilities found in separated flows around bluff bodies that controlled actuation should target at are briefly presented. Key parameters for effective control are introduced using the nominally two-dimensional flow around a circular cylinder as a paradigm. The effects of the actuator configuration and location, amplitude and frequency of excitation, input waveform, as well as the phase difference between individual actuators are illustrated through examples classified based on symmetry properties. In general, symmetric excitation at frequencies higher than approximately five times the uncontrolled frequency of vortex shedding acts destructively on regular vortex shedding and can be safely employed for reducing the mean drag and lift fluctuations. Antisymmetric and symmetric excitation at low frequencies of the order of the natural frequency can amplify the wake instability and increase the mean and fluctuating aerodynamic forces, respectively, due to vortex locking-on to the excitation frequency or its subharmonics. Results from several studies show that the geometry and arrangement of the electrodes is of utmost significance. Power consumption is typically very low, but the electromechanical efficiency can be optimized by input waveform modulation.