Faster Identification Time Research Articles

A fast damping ratio identification method for rotating machinery based on the singular value of a directional power spectra density function matrix

It is of great importance to accurately evaluate the stability margin of a turbomachine both in shop tests and in operation. The damping ratio of the first forward whirl mode is the key indicator of rotordynamic stability. In this paper, a fast operational modal analysis (OMA) technique based on the singular value of the directional power spectra density function (dPSDF) matrix was developed to estimate the damping ratios of the first forward and backward whirl modes. Numerical simulations and experiments were carried out to verify the accuracy of this method on a high-speed test rig. Moreover, the dPSDF-based method was compared to two conventional OMA methods - data-driven and covariance-driven stochastic subspace identification (SSI) methods - resulting in more stable and accurate estimates for the damping ratio along with a much faster identification time (less than 0.5 s). This study provides an accurate and fast method for performing stability shop tests for turbo machines and stability online monitoring.

Speaker identification using multimodal neural networks and wavelet analysis

The rapid momentum of the technology progress in the recent years has led to a tremendous rise in the use of biometric authentication systems. The objective of this research is to investigate the problem of identifying a speaker from its voice regardless of the content. In this study, the authors designed and implemented a novel text-independent multimodal speaker identification system based on wavelet analysis and neural networks. Wavelet analysis comprises discrete wavelet transform, wavelet packet transform, wavelet sub-band coding and Mel-frequency cepstral coefficients (MFCCs). The learning module comprises general regressive, probabilistic and radial basis function neural networks, forming decisions through a majority voting scheme. The system was found to be competitive and it improved the identification rate by 15% as compared with the classical MFCC. In addition, it reduced the identification time by 40% as compared with the back-propagation neural network, Gaussian mixture model and principal component analysis. Performance tests conducted using the GRID database corpora have shown that this approach has faster identification time and greater accuracy compared with traditional approaches, and it is applicable to real-time, text-independent speaker identification systems.

Optical penetration-based silkworm pupa gender sensor structure

This paper proposes and experimentally demonstrates for what is believed to be the first time a highly sought-after optical structure for highly-accurate identification of the silkworm pupa gender. The key idea is to exploit a long wavelength optical beam in the red or near infrared spectrum that can effectively and safely penetrate the body of a silkworm pupa. Later on, simple image processing operations via image thresholding, blob filtering, and image inversion processes are applied in order to eliminate the unwanted image noises and at the same time highlight the gender gland. Experimental proof of concept using three 636 nm wavelength light emitting diodes, a two-dimensional web camera, an 8 bit microcontroller board, and a notebook computer shows a very high 95.6% total accuracy in identifying the gender of 45 silkworm pupae with a measured fast identification time of 96.6 ms. Other key features include low cost, low component counts, and ease of implementation and control.

Demonstration of a single-wavelength spectral-imaging-based Thai jasmine rice identification

A single-wavelength spectral-imaging-based Thai jasmine rice breed identification is demonstrated. Our nondestructive identification approach relies on a combination of fluorescent imaging and simple image processing techniques. Especially, we apply simple image thresholding, blob filtering, and image subtracting processes to either a 545 or a 575 nm image in order to identify our desired Thai jasmine rice breed from others. Other key advantages include no waste product and fast identification time. In our demonstration, UVC light is used as our exciting light, a liquid crystal tunable optical filter is used as our wavelength seclector, and a digital camera with 640 active pixels × 480 active pixels is used to capture the desired spectral image. Eight Thai rice breeds having similar size and shape are tested. Our experimental proof of concept shows that by suitably applying image thresholding, blob filtering, and image subtracting processes to the selected fluorescent image, the Thai jasmine rice breed can be identified with measured false acceptance rates of <22.9% and <25.7% for spectral images at 545 and 575 nm wavelengths, respectively. A measured fast identification time is 25 ms, showing high potential for real-time applications.

A Combined Polling and Random Access Technique for Enhanced Anti-Collision Performance in RFID Systems

In this paper we propose a novel RFID anti-collision technique that intelligently combines polling and random access schemes. These two fundamentally different medium access control protocols are coherently integrated in our design while functionally complementing each other. The polling mode is designed to enable fast collision-free identification for the tags that exist within reader's coverage across the sessions. In contrast, the random access mode attempts to read the tags uncovered by the polling mode. Our proposed technique is particularly suited for a class of RFID applications in which a stationary reader periodically attempts to identify the tags with slow mobility. Numerical results show that our proposed technique yields much faster identification time against the existing approaches under various operating conditions.