Discrete interferences optimum beamformer in correlated signal and interfering noise

This paper introduces a significant special situation where the noise is a collection of D-plane interference signals and the correlated noise of D+1 is less than the number of array components. An optimal beamforming processor based on the minimum variance distortionless response (MVDR) generates and combines appropriate statistics for the D+1 model. Instead of the original space of the N-dimensional problem, the interference signal subspace is reduced to D+1. Typical antenna arrays in many modern communication networks absorb waves generated from multiple point sources. An analytical formula was derived to improve the signal to interference and noise ratio (SINR) obtained from the steering errors of the two beamformers. The proposed MVDR processor-based beamforming does not enforce general constraints. Therefore, it can also be used in systems where the steering vector is compromised by gain. Simulation results show that the output of the proposed beamformer based on the MVDR processor is usually close to the ideal state within a wide range of signal-to-noise ratio and signal-to-interference ratio. The MVDR processor-based beamformer has been experimentally evaluated. The proposed processor-based MVDR system significantly improves performance for large interference white noise ratio (INR) in the sidelobe region and provide an appropriate beam pattern.

Stochastic dynamics of chemotactic colonies with logistic growth

The interplay between cellular growth and cell-cell signaling is essential for the aggregation and proliferation of bacterial colonies, as well as for the self-organization of cell tissues. To investigate this interplay, we focus here on the collective properties of dividing chemotactic cell colonies by studying their long-time and large-scale dynamics through a renormalization group (RG) approach. The RG analysis reveals that a relevant but unconventional chemotactic interaction -- corresponding to a polarity-induced mechanism -- is generated by fluctuations at macroscopic scales, even when an underlying mechanism is absent at the microscopic level. This emerges from the interplay of the well-known Keller--Segel (KS) chemotactic nonlinearity and cell birth and death processes. At one-loop order, we find no stable fixed point of the RG flow equations. We discuss a connection between the dynamics investigated here and the celebrated Kardar--Parisi--Zhang (KPZ) equation with long-range correlated noise, which points at the existence of a strong-coupling, nonperturbative fixed point.

Structure diagram and dynamics of formation of hexagonal boron nitride in phase-field crystal model

The phase-field crystal (PFC-model) is a powerful tool for modelling of the crystallization in colloidal and metallic systems. In the present work, the modified hyperbolic phase-field crystal model for binary systems is presented. This model takes into account slow and fast dynamics of moving interfaces for both concentration and relative atomic number density (which were taken as order parameters). The model also includes specific mobilities for each dynamical field and correlated noise terms. The dynamics of chemical segregation with origination of mixed pseudo-hexagonal binary phase (the so-called ‘triangle phase’) is used as a benchmark in two spatial dimensions for the developing model. Using the free energy functional and specific lattice vectors for hexagonal crystal, the structure diagram of co-existence of liquid and three-dimensional hexagonal phase for the binary PFC-model was carried out. Parameters of the crystal lattice correspond to the hexagonal boron nitride (BN) crystal, the values of which have been taken from the literature. The paper shows the qualitative agreement between the developed structure diagram of the PFC model and the previously known equilibrium diagram for BN constructed using thermodynamic functions. This article is part of the theme issue ‘Transport phenomena in complex systems (part 2)’.

Uncertainty Assessment for HATPRO Microwave Radiometer Measurements and Calibrations

<p>The atmospheric boundary layer (ABL) is the most important under-sampled part of the atmosphere. ABL monitoring is crucial for short-range forecasting of severe weather within highly resolving numerical weather predictions (NWP). Top-priority atmospheric variables for NWP applications like temperature (T) and humidity (H) profiles are currently not adequately measured. Ground-based microwave radiometers (MWRs) like HATPRO (Humidity And Temperature PROfiler) are particularly well suited to obtain such T-profiles in the ABL as well as coarse resolution H-profiles. It has been shown by previous studies that the assimilation of MWR observations is beneficial for NWP models, however MWR data are not yet routinely assimilated into operational NWP. The HATPRO measures in zenith and other angles throughout the troposphere over an area with ~10 km radius and has a temporal resolution on the order of seconds. Measured brightness temperatures (TB) are used to retrieve the T- and H-profiles. Path integrated values IWV (Integrated Water Vapor) and LWP (Liquid Water Path) are quite reliable with excellent uncertainties up to 0.5 kg/m<sup>2</sup> and 20 g/m<sup>2</sup>, respectively.</p> <p>Driven by the E-PROFILE program, a business case proposal was recently accepted by EUMETNET to continuously provide MWR data to the European meteorological services. Also, the European Research Infrastructure for the observation of Aerosol, Clouds, and Trace gases (ACTRIS) and the European COST action PROBE (PROfiling the atmospheric Boundary layer at European scale) currently focus on establishing continent-wide quality and observation standards for MWR networks for research as well as for NWP applications. The German Weather Service (DWD) also investigates the potential of HATPRO networks for improving short-term weather forecasts over Germany.</p> <p>For all this it is important to obtain an overview of what HATPROs are capable of in regard to their measurement uncertainty. This was done by conducting coordinated experiments at JOYCE (Jülich Observatory for Cloud Evolution) and the FESSTVaL (Field Experiment on Submesoscale Spatio-Temporal Variability at Lindenberg) campaign in 2021 within a prototype MWR network. The goal is to develop a standard procedure for error characterization that can be applied to any HATPRO network instrument (guidance for operators).</p> <p>Important error components are absolute calibration errors (biases), drifts (instrument stability, leaps between calibrations), radiometric noise and also location specific radio frequency interferences (RFI). For the absolute calibration with liquid nitrogen, the repeatability, the integration time and the time between calibrations are essential. Differences between consecutive calibrations are analysed, the right duration of a calibration and the right amount of time between calibrations are proposed, referring to the magnitude of the observed drifts. For the determination of noise levels for each channel, covariance matrices (correlated noise) of measured brightness temperatures on the cold- and hotload references are presented. RFI are detectable via clear-sky azimuth- and/or elevation scans.</p>

Noise-trained deep neural networks effectively predict human vision and its neural responses to challenging images

Deep neural networks (DNNs) for object classification have been argued to provide the most promising model of the visual system, accompanied by claims that they have attained or even surpassed human-level performance. Here, we evaluated whether DNNs provide a viable model of human vision when tested with challenging noisy images of objects, sometimes presented at the very limits of visibility. We show that popular state-of-the-art DNNs perform in a qualitatively different manner than humans—they are unusually susceptible to spatially uncorrelated white noise and less impaired by spatially correlated noise. We implemented a noise training procedure to determine whether noise-trained DNNs exhibit more robust responses that better match human behavioral and neural performance. We found that noise-trained DNNs provide a better qualitative match to human performance; moreover, they reliably predict human recognition thresholds on an image-by-image basis. Functional neuroimaging revealed that noise-trained DNNs provide a better correspondence to the pattern-specific neural representations found in both early visual areas and high-level object areas. A layer-specific analysis of the DNNs indicated that noise training led to broad-ranging modifications throughout the network, with greater benefits of noise robustness accruing in progressively higher layers. Our findings demonstrate that noise-trained DNNs provide a viable model to account for human behavioral and neural responses to objects in challenging noisy viewing conditions. Further, they suggest that robustness to noise may be acquired through a process of visual learning.