Anindya Lal Roy

Capturing Complementary Information via Reversed Filter Bank and Parallel Implementation with MFCC for Improved Text-Independent Speaker Identification

A state of the art speaker identification (SI) system requires a robust feature extraction unit followed by a speaker modeling scheme for generalized representation of these features. Over the years, mel-frequency cepstral coefficients (MFCC) modeled on the human auditory system have been used as a standard acoustic feature set for SI applications. However, due to the structure of its filter bank, it captures vocal tract characteristics more effectively in the lower frequency regions. This work proposes a new set of features using a complementary filter bank structure which improves distinguishability of speaker specific cues present in the higher frequency zone. Unlike high level features that are difficult to extract, the proposed feature set involves little computational burden during the extraction process. When combined with MFCC via a parallel implementation of speaker models, the proposed feature improves performance baseline of MFCC based system. The proposition is validated by experiments conducted on two different kinds of databases namely YOHO (microphone speech) and POLYCOST (telephone speech) with two different classifier paradigms, namely Gaussian Mixture Models (GMM) and Polynomial Classifier (PC) and for various model orders

Fusion of a Complementary Feature Set with MFCC for Improved Closed Set Text-Independent Speaker Identification

A state of the art speaker identification (SI) system requires a robust feature extraction unit followed by a speaker modeling scheme for generalized representation of these features. Over the years, Mel-frequency cepstral coefficients (MFCC) modeled on the human auditory system have been used as a standard acoustic feature set for SI applications. However, due to the structure of its filter bank, it captures vocal tract characteristics more effectively in the lower frequency regions. This work proposes a new set of features using a complementary filter bank structure which improves distinguishability of speaker specific cues present in the higher frequency zone. Unlike high level features that are difficult to extract, the proposed feature set involves little computational burden during the extraction process. When combined with MFCC via a parallel implementation of speaker models, the proposed feature improves performance baseline of MFCC based system. The proposition is validated by experiments conducted on two different kinds of databases namely YOHO (microphone speech) and POLYCOST (telephone speech) with Gaussian mixture model (GMM) as a classifier for various model orders.

Analysis of the Pull-In Phenomenon in Microelectromechanical Varactors

High-Q factor voltage controlled oscillators (VCO) need a wide tuning range and low phase noise over gigahertz ranges of frequency which depends on the tunability of the capacitors in the LC tank circuit. The reasons behind the development of a micro electro mechanical (MEM) varactor were the difficulties encountered in the realization of on-chip variable capacitors having low phase noise and high quality factors with a wide tuning range in the span of frequencies over process and temperature variations. This paper presents an efficient closed-form model for determination of the pull-in voltage in a surface micro machined MEM varactor which is a factor directly affecting the tunability of the device. The nonlinear spring hardening effects associated with proper load-deflection characteristics of clamped plates and the electrostatic spring softening effects due to the parallel-plate and fringing field capacitances have been taken into account with the dimensions of the device optimized through finite element analysis (FEA).

MEMS Piezoresistive Accelerometers

MEMS piezoresistive accelerometers are inertial sensors which measure acceleration of the reference frame to which they are attached. These devices provide extremely localized acceleration-induced stress sensing with low noise outputs and have been the subject of academic as well as commercial research for quite a few years. This chapter deals with the basics micromachined piezoresistive accelerometers, tracing their evolution and typical analyses to sensor fabrication and characterization.

Low-Field Transport in Bovine Serum Albumin Conjugated ZnO Nanoparticle Networks

A Au/bovine serum albumin (BSA)-ZnO (n-type nanoparticles)/Au system is investigated for its transport properties using electrical (I-V) characterization in conjunction with spectroscopy techniques for cost-effective power-efficient biosensing. Temperature-dependent variation in different conduction mechanisms is observed with Au/ZnO/Au system as reference. Schottky emission dominates the carrier transport with defect-assisted tunneling and Poole-Frenkel hopping contributing peripherally. The distribution of defect states within the ZnO bandgap appears to be responsible for such low-field conduction characteristics along with the surface dipole effect of the BSA, which modulates barrier height and, consequently, carrier population in the ZnO nanoparticles.

Symmetry breaking oscillations in electrostatic MEMS under superharmonic excitation

Microelectromechanical systems (MEMS)-based vibratory structures have become popular due to their simple fabrication processes, high quality factors, low loss and high sensitivity. Applications of such devices as sensors require a proper understanding of their operational characteristics in various regimes. This work is a study of the dynamics of an electrostatically actuated MEMS doubly-clamped beam under superharmonic excitation from a qualitative point of view whereby the nonlinear dynamic response of such devices can be analyzed to some extent. Here, we report some phenomena which we have come across during our measurements of the MEMS device under consideration and attempt to explain our findings through some intuitive analysis.

Small-deflection analysis of microelectromechanical varactors

Voltage controlled oscillators (VCO) with high quality (Q) factors require wide tuning ranges and low phase noise performance which depends on the tunability of the capacitors in the LC tank circuit. Microelectromechanical (MEM) varactors are being considered as replacements for VLSI on-chip variable capacitors due to the susceptibility of the latter to phase noise and low Q-factors. This paper presents a theoretical model for the determination of the pull-in voltage in a surface micromachined MEM varactor which is a factor directly affecting the tunability of the device. The associated electrostatic spring softening effects due to the parallel-plate and fringing field capacitances have been taken into account in developing the semi-analytical model with extensive finite element method (FEM) and lumped parameter simulations being carried out.