RESUMO
In this paper, we investigate different scenarios of anomaly detection on decentralised Internet of Things (IoT) applications. Specifically, an anomaly detector is devised to detect different types of anomalies for an IoT data management system, based on the decentralised alternating direction method of multipliers (ADMM), which was proposed in our previous work. The anomaly detector only requires limited information from the IoT system, and can be operated using both a mathematical-rule-based approach and the deep learning approach proposed in the paper. Our experimental results show that detection based on mathematical approach is simple to implement, but it also comes with lower detection accuracy (78.88%). In contrast, the deep-learning-enabled approach can easily achieve a higher detection accuracy (96.28%) in the real world working environment.
RESUMO
The present work assessed Mandarin sentence understanding when the electric and acoustic portions are not temporally aligned in simulated combined electric-and-acoustic stimulation (EAS). A relative time shift was added between the electric and acoustic portions, simulating the temporal misalignment effect in EAS processing. The processed stimuli were played to normal-hearing listeners to recognize. Experimental results showed a significant decrease of the intelligibility score caused by the temporal misalignment in the two portions of EAS processing, suggesting the need to avoid temporal misalignment in EAS. The preceding acoustic-portion more significantly decreased the understanding of EAS-processed Mandarin stimuli than the preceding electric-portion.
RESUMO
Hearing loss severely affects human speech communication and the quality of life, and efficient hearing screening can help hearing function diagnosis and subsequent hearing rehabilitation. Pure-tone audiometry test and distortion-product otoacoustic emission (DPOAE) measurement are two commonly-used clinical techniques for hearing loss diagnosis, and they were developed based on different mechanisms in a hearing process. Early work investigated the relation between audiometric thresholds and DPOAE measurements. The present work aimed to use a simple linear fitting to estimate audiometric pure-tone threshold averages (PTAs) from DPOAE signal-to-noise-ratio (SNR) measurements, i.e., DPOAE amplitude minus the mean noise floor. Audiometric PTA values and DPOAE SNRs were measured from both ears of 30 listeners with normal hearing or mild-to-moderate hearing loss. The DPOAE SNR measurements of 4 distortion products (i.e., 2fl-f2= 1, 2, 4 and 8 kHz) were combined with a linear prediction model, and correlated with the PTA values. Data analysis showed moderate high correlation coefficients (i.e., r=0.84 and 0.64) for the left and right ears, respectively. The results of the present work demonstrate the possibility to estimate the behavioral audiometric PTA values from the objective DPOAE SNR measurements for hearing loss diagnosis.
Assuntos
Surdez , Qualidade de Vida , Audiometria de Tons Puros , Cóclea , Audição , Humanos , Emissões Otoacústicas EspontâneasRESUMO
Accurate gait event detection is necessary for control strategies of gait rehabilitation robots. However, due to personal diversity between individuals, it is a challenge for robots to detect a gait event at various stride frequencies. This paper proposes a novel method for gait event detection of a gait rehabilitation robot using a single inertial sensor mounted on the thigh. A self-adaptive threshold for detecting heel strike is obtained in real time via a linear regression model. Observable thresholds for toe off detection are constant at various stride frequencies. Experiments are conducted based on 20 healthy subjects and six hemiplegic patients wearing a gait rehabilitation robot and walking at various kinds of stride frequencies. The experimental results show that the proposed method can detect heel strike and toe off gait events within an average 2% gait cycle temporal errors and never miss two-gait event detection. Compared to the conventional thresholding method, this work presents a simple and robust application for gait event detection in healthy and hemiplegic subjects by one inertial sensor. The linear regression model can be applicable to different subjects walking at various stride frequencies.
RESUMO
Secondary phases are common in Cu2ZnSnS4 (CZTS) thin films, which can be fatal to the performance of solar cell devices fabricated from this material. They are difficult to detect by X-Ray diffraction (XRD) because of the weak peak in spectra compared with the CZTS layer. Herein, it was found that in-depth elemental distribution by a secondary ion mass spectroscopy method illustrated uniform film composition in the bulk with slight fluctuation between different grains. X-ray photoelectron spectroscopy (XPS) measurement was conducted after sputtering the layer with different depths. An Auger electron spectrum with Auger parameter were used to check the chemical states of elements and examine the distribution of secondary phases in the CZTS films. Secondary phases of CuS, ZnS and SnS were detected at the surface of the CZTS film within a 50-nm thickness while no secondary phases were discovered in the bulk. The solar cell fabricated with the as-grown CZTS films showed a conversion efficiency of 2.1% (Voc: 514.3 mV, Jsc: 10.4 mA/cm2, FF: 39.3%) with an area of 0.2 cm2 under a 100 mW/cm2 illumination. After a 50-nm sputtering on the CZTS film, the conversion efficiency of the solar cell was improved to 6.2% (Voc: 634.0 mV, Jsc: 17.3 mA/cm2, FF: 56.9%).