A Deep Learning Framework for Evaluating the Over-the-Air Performance of the Antenna in Mobile Terminals.
Research
RF Safe Research Library
Jan 1, 2024
This engineering study presents RTEEMF-PhoneAnts, a convolutional neural network framework to predict mobile phone antenna metrics (EIRP, TRP, and SAR) from near-field EMF distributions. Using a dataset of 7000 mobile phone models, the authors report statistical validation for predicted SAR/TRP and a validation…
Conformational changes of β-thalassemia major hemoglobin and oxidative status of plasma after in vitro exposure to extremely low-frequency electromagnetic fields: An artificial neural network analysis.
Research
RF Safe Research Library
Jan 1, 2021
This in vitro study exposed blood samples from β-thalassemia major patients and matched healthy donors (n=12 each) to 50 Hz ELF-EMF at 0.5 and 1 mT for 1–2 hours. The authors report that two-hour exposure induced significant oxidative changes in β-thalassemia major samples and examined protein/Hb structural features…
Artificial Neural Network-Based Uplink Power Prediction From Multi-Floor Indoor Measurement Campaigns in 4G Networks.
Research
RF Safe Research Library
Jan 1, 2021
This paper describes multi-floor indoor LTE measurement campaigns aimed at enabling assessment and monitoring of RF-EMF exposure, with emphasis on uplink exposure via mobile phone transmit power. Because uplink transmit power is not directly available from off-the-shelf devices, the authors propose an artificial…
Weak electric fields detectability in a noisy neural network.
Research
RF Safe Research Library
Jan 1, 2017
This modeling study examines how a noisy spiking neural network detects a weak applied electric field using an Izhikevich neuron framework with excitatory/inhibitory neurons and conduction delays. The authors report stochastic resonance driven by white noise that enhances detectability, and that this effect largely…
Advancing profiling sensors with a wireless approach.
Research
RF Safe Research Library
Jan 1, 2012
This engineering paper describes a wireless implementation of a near-infrared profiling sensor using multiple sensor nodes and a base station for data preprocessing and realignment. It reports an enhanced classification approach using a back-propagation neural network to categorize objects as human, animal, or…