arXiv subjects

Rohit Kumar

Publications and source records attributed to Rohit Kumar.

104 records · Page 6Linked to original sources

arXiv2019

Unsupervised Neural Mask Estimator For Generalized Eigen-Value Beamforming Based ASR

The state-of-art methods for acoustic beamforming in multi-channel ASR are based on a neural mask estimator that predicts the presence of speech and noise. These models are trained using a paired corpus of clean and noisy recordings (teacher model). In this paper, we attempt to move away from the requirements of having supervised clean recordings for training the mask estimator. The models based on signal enhancement and beamforming using multi-channel linear prediction serve as the required mask estimate. In this way, the model training can also be carried out on real recordings of noisy speech rather than simulated ones alone done in a typical teacher model. Several experiments performed on noisy and reverberant environments in the CHiME-3 corpus as well as the REVERB challenge corpus highlight the effectiveness of the proposed approach. The ASR results for the proposed approach provide performances that are significantly better than a teacher model trained on an out-of-domain dataset and on par with the oracle mask estimators trained on the in-domain dataset.

eess.AS↗

arXiv2018

Learning to Coordinate in a Decentralized Cognitive Radio Network in Presence of Jammers

Efficient utilization of licensed spectrum in the cognitive radio network is challenging due to lack of coordination among the Secondary Users (SUs). Distributed algorithms proposed in the literature aim to maximize the network throughput by ensuring orthogonal channel allocation for the SUs. However, these algorithms work under the assumption that all the SUs faithfully follow the algorithms which may not always hold due to the decentralized nature of the network. In this paper, we study distributed algorithms that are robust against malicious behavior (jamming attack). We consider both the cases of jammers launching coordinated and uncoordinated attacks. In the coordinated attack, the jammers select non-overlapping channels to attack in each time slot and can significantly increase the number of collisions for SUs. We setup the problem in each scenario as a multi-player bandit and develop algorithms. The analysis shows that when the SUs faithfully implement proposed algorithms, the regret is constant with high probability. We validate our claims through exhaustive synthetic experiments and also through a realistic USRP based experiments.

eess.SP↗

arXiv2018

Distributed Learning Algorithms for Opportunistic Spectrum Access in Infrastructure-less Networks

An opportunistic spectrum access (OSA) for the infrastructure-less (or cognitive ad-hoc) network has received significant attention thanks to emerging paradigms such as the Internet of Things (IoTs) and smart grids. Research in this area has evolved from the \r{ho}rand algorithm requiring prior knowledge of the number of active secondary users (SUs) to the musical chair (MC) algorithm where the number of SUs are unknown and estimated independently at each SU. These works ignore the number of collisions in the network leading to wastage of power and bring down the effective life of battery operated SUs. In this paper, we develop algorithms for OSA that learn faster and incurs fewer number of collisions i.e. energy efficient. We consider two types of infrastructure-less decentralized networks: 1) static network where the number of SUs are fixed but unknown, and 2) dynamic network where SUs can independently enter or leave the network. We set up the problem as a multi-player mult-armed bandit and develop two distributed algorithms. The analysis shows that when all the SUs independently implement the proposed algorithms, the loss in throughput compared to the optimal throughput, i.e. regret, is a constant with high probability and significantly outperforms existing algorithms both in terms of regret and number of collisions. Fewer collisions make them ideally suitable for battery operated SU terminals. We validate our claims through exhaustive simulated experiments as well as through a realistic USRP based experiments in a real radio environment.

eess.SP↗

arXiv2018

Throughput Optimized Non-Contiguous Wideband Spectrum Sensing via Online Learning and Sub-Nyquist Sampling

In this paper, we consider non-contiguous wideband spectrum sensing (WSS) for spectrum characterization and allocation in next generation heterogeneous networks. The proposed WSS consists of sub-Nyquist sampling and digital reconstruction to sense multiple non-contiguous frequency bands. Since the throughput (i.e. the number of vacant bands) increases while the probability of successful reconstruction decreases with increase in the number of sensed bands, we develop an online learning algorithm to characterize and select frequency bands based on their spectrum statistics. We guarantee that the proposed algorithm allows sensing of maximum possible number of frequency bands and hence, it is referred to as throughput optimized WSS. We also provide a lower bound on the number of time slots required to characterize spectrum statistics. Simulation and experimental results in the real radio environment show that the performance of the proposed approach converges to that of Myopic approach which has prior knowledge of spectrum statistics.

eess.SP↗

arXiv2018

Magnetotransport studies of optimally doped Sr(Fe${_{1-x}}$Co${_x}$)${_2}$As${_2}$

We report magnetotransport measurements and its scaling analysis for the optimally electron doped Sr(Fe${_{0.88}}$Co${_{0.12}}$)${_2}$As${_2}$ system. We pbserve that both the Kohler's and modified Kohler's scalings are violated. Interestingly, the Hall angle displays a quadratic temperature dependence similar to many cuprates and heavy fermion systems. The fact that this temperature dependence is seen in spite of the violation of modified Kohler's scaling suggests that the Hall angle and the magnetoresistance are not governed by the same scattering mechanism. We also observe a linear magnetoresistance in this system, which does not harbor a spin density wave ground state. Implcations of our observations are discussed in the context of spin fluctuations in strongly correlated electron systems.

cond-mat.supr-con↗

arXiv2018

Production of sunspots and their effects on the corona and solar wind: Insights from a new 3D flux-transport dynamo model

We present a three-dimensional numerical model for the generation and evolution of the magnetic field in the solar convection zone, in which sunspots are produced and contribute to the cyclic reversal of the large-scale magnetic field. We then assess the impact of this dynamo-generated field on the structure of the solar corona and solar wind. This model solves the induction equation in which the velocity field is prescribed. This velocity field is a combination of a solar-like differential rotation and meridional circulation. We develop an algorithm that enables the magnetic flux produced in the interior to be buoyantly transported towards the surface to produce bipolar spots. We find that those tilted bipolar magnetic regions contain a sufficient amount of flux to periodically reverse the polar magnetic field and sustain dynamo action. We then track the evolution of these magnetic features at the surface during a few consecutive magnetic cycles and analyze their effects on the topology of the corona and on properties of the solar wind (distribution of streamers and coronal holes, and of slow and fast wind streams) in connection with current observations of the Sun.

astro-ph.SR↗

arXiv2018

High temperature linear magnetoresistance and scaling behavior in the Ba(Fe${_{1-x}}$Co${_{x}}$)$_{2}$As$_{2}$ series

We present magnetotransport studies of the parent, an underdoped and an optimally doped composition of the Ba(Fe${_{1-x}}$Co${_{x}}$)${_{2}}$As${_{2}}$ series. We observe that both the Kohler's and modified Kohler's scaling is typically violated in both the magnetically ordered and paramagnetic regimes. A notable exception is the magnetically ordered state of the underdoped composition where the modified Kohler's scaling is observed, indicating its relative similarity to the cuprates and some heavy fermion systems. This composition also exhibits a feature in the Hall angle, which could signify the opening of a pseudogap before the onset of long range magnetic order. Interestingly, the transverse magnetoresistance is seen to exhibit a linear field dependence in the paramagnetic regimes of all these compositions. We also demonstrate that the $B/T$ scaling proposed recently in the context of quantum critical systems is seen to be valid in all these systems. The implications of our observations are discussed in the context of magnetotransport of metals with incipient magnetic fluctuations.

cond-mat.supr-con↗

arXiv2017

Magnetic disorder and gap symmetry in optimally electron doped Sr(Fe, Co)$_2$As$_2$ superconductor

We investigate the magnetic pair-breaking due to Mn impurities in the optimally electron doped Sr(Fe$_{0.88} $Co$_{0.12} $)$_{2} $As$_2 $ superconductor to deduce the symmetry of the superconducting order parameter. Experiments on the as-grown crystals reveal a T$_c $ suppression rate of $ \sim $30 mK/$μΩcm$, which is in close agreement with similarly slower values of T$_c $ suppression rates reported previously for various transition metal impurities, both, magnetic and non-magnetic, in several structurally analogous iron-based superconductors. However, careful annealing of these crystals at low temperature for longer durations reveals new information crucial to the determination of the pairing symmetry. We found that the crystallographic defects are a significant source of pair-breaking in the as-grown crystals. We first establish that these defects are point-like by showing that their sole effect on electrical transport is to add a temperature independent scattering term that shifts the whole $ ρ$ vs. T curves rigidly up. The T$_c $ suppression rate due to these point-like defects is slow, $ \le$ 35 mK/$μΩcm$. On the other hand, T$_c $ suppression rate due to magnetic pair-breaking is estimated to be faster than 325 mK/$μΩcm$. A slower pair-breaking rate (measured in mK/$μΩcm$) than expected due to non-magnetic crystallographic defects, together with a faster pair-breaking rate due to magnetic impurities disfavors a sign-changing $s_{+-}$-wave and argues in the favor of a non-sign-changing $s_{++}$-wave state in the optimally electron doped SrFe$_2 $As$_2 $ superconductor.

cond-mat.supr-con↗

arXiv2017

Amplification of large-scale magnetic field in nonhelical magnetohydrodynamics

It is typically assumed that the kinetic and magnetic helicities play a crucial role in the growth of large-scale dynamo. In this paper we demonstrate that helicity is not essential for the amplification of large-scale magnetic field. For this purpose, we perform nonhelical magnetohydrodynamic (MHD) simulation, and show that the large-scale magnetic field can grow in nonhelical MHD when random external forcing is employed at scale $1/10$ the box size. The energy fluxes and shell-to-shell transfer rates computed using the numerical data show that the large-scale magnetic energy grows due to the energy transfers from the velocity field at the forcing scales.

physics.flu-dyn↗

arXiv2017

Dynamo transition in a five-mode helical model

We construct a five-mode helical dynamo model containing three velocity and two magnetic modes and solve it analytically. This model exhibits dynamo transition via supercritical pitchfork bifurcation. We show that the critical magnetic Reynolds number for dynamo transition ($\mathrm{Rm}_c$) asymptotes to constant values for very low and very high magnetic Prandtl numbers ($\mathrm{Pm}$). Beyond dynamo transition, secondary bifurcations lead to periodic, quasi-periodic, and chaotic dynamo states as the forcing amplitude is increased and chaos appears through a quasi-periodic route.

physics.flu-dyn↗

arXiv2016

Dynamos at extreme magnetic Prandtl numbers: Insights from shell models

We present an MHD shell model suitable for computation of various energy fluxes of magnetohydrodynamic turbulence for very small and very large magnetic Prandtl numbers $\mathrm{Pm}$; such computations are inaccessible to direct numerical simulations. For small $\mathrm{Pm}$, we observe that both kinetic and magnetic energy spectra scale as $k^{-5/3}$ in the inertial range, but the dissipative magnetic energy scales as $k^{-11/3}\exp(-k/k_η)$. Here, the kinetic energy at large length scale feeds the large-scale magnetic field that cascades to small-scale magnetic field, which gets dissipated by Joule heating. The large-$\mathrm{Pm}$ dynamo has a similar behaviour except that the dissipative kinetic energy scales as $k^{-13/3}$. For this case, the large-scale velocity field transfers energy to the large-scale magnetic field, which gets transferred to small-scale velocity and magnetic fields; the energy of the small-scale magnetic field also gets transferred to the small-scale velocity field, and the energy thus accumulated is dissipated by the viscous force.

physics.flu-dyn↗

arXiv2015

Energy transfers in dynamos with small magnetic Prandtl numbers

We perform numerical simulation of dynamo with magnetic Prandtl number $\mathrm{Pm} =0.2$ on $1024^3$ grid, and compute the energy fluxes and the shell-to-shell energy transfers. These computations indicate that the magnetic energy growth takes place mainly due to the energy transfers from large-scale velocity field to large-scale magnetic field and that the magnetic energy flux is forward. The steady-state magnetic energy is much smaller than the kinetic energy, rather than equipartition; this is because the magnetic Reynolds number is near the dynamo transition regime. We also contrast our results with those for dynamo with $\mathrm{Pm} =20$ and decaying dynamo.

physics.flu-dyn↗

arXiv2014

Energy transfers and magnetic energy growth in small-scale dynamo

In this letter we investigate the dynamics of magnetic energy growth in small-scale dynamo by studying energy transfers, mainly energy fluxes and shell-to-shell energy transfers. We perform dynamo simulations for magnetic Prandtl number $\mathrm{Pm}=20$ on $1024^3$ grid using pseudospectral method. We demonstrate that the magnetic energy growth is caused by nonlocal energy transfers from the large-scale or forcing-scale velocity field to small-scale magnetic field. The peak of these energy transfers move towards lower wavenumbers as dynamo evolves, which is the reason why the integral scale of the magnetic field increases with time. The energy transfers $U2U$ (velocity to velocity) and $B2B$ (magnetic to magnetic) are forward and local.

physics.flu-dyn↗

arXiv2013

Dynamo in Protostar

In this paper, we estimate the magnetic Reynolds number of a typical protostar before and after deuterium burning, and claim for the existence of dynamo process in both the phases, because the magnetic Reynolds number of the protostar far exceeds the critical magnetic Reynolds number for dynamo action. Using the equipartition of kinetic and magnetic energies, we estimate the steady-state magnetic field of the protostar to be of the order of kilo-gauss, which is in good agreement with observations.

astro-ph.SR↗