Modeling and Mitigation of 7.125-7.40 GHz Terrestrial Network RFI on the Passive Earth Exploration Satellite Service in 6.725-7.125 GHz Band

The 7.125-7.4 GHz band is attractive for next generation Terrestrial Network (TN) deployments owing to the large bandwidths available and favorable propagation characteristics. Furthermore, recent U.S. Presidential actions have cleared the usage of this band for 6G by stipulating relocation of federal incumbents that share this band. However, this deployment can only be successful if we can also guarantee coexistence of these networks with existing incumbents operating in adjacent bands. This paper presents a comprehensive analysis of the Radio Frequency Interference (RFI) caused by the proposed TNs in the 7.125-7.4 GHz band at passive Earth Exploration Satellite Service (EESS) sensors that operate in the adjacent 6.725-7.125 GHz band. Using TN base stations (BSs) equipped with filtennas (filtering antennas) as well as transmit precoders for RFI mitigation, we carry out an RFI analysis that accounts for increasing BS deployments in the contiguous U.S. over a 10 year period from 2030 to 2040. We also characterize the size of the guard bands needed to protect the EESS sensors for different BS deployment densities. With appropriate filtenna and precoder design, our results reveal that a 100 Mbps increase in the rate requirements of the TN users results in an RFI increase of roughly 2.45 dB at the EESS sensors. For a 25 MHz Guard Band, simulations show that in 2030, there is no significant RFI for user rates upto 500 Mbps. However, the same systems in 2040 would cause RFI that is around 4 dB above the ITU mandated threshold for passive EESS sensors. This would need to be countered by (a) increasing Guard Band width to 35 MHz, or (b) by reducing the user data rate requirements to 300 Mbps.

ResEMGNet: A Lightweight Residual Deep Learning Architecture for Neuromuscular Disorder Detection from Raw EMG Signals

Amyotrophic Lateral Sclerosis (ALS) and Myopathy are debilitating neuromuscular disorders that demand accurate and efficient diagnostic approaches. In this study, we harness the power of deep learning techniques to detect ALS and Myopathy. Convolutional Neural Networks (CNNs) have emerged as powerful tools in this context. We present ResEMGNet, designed to identify ALS and Myopathy directly from raw electromyography (EMG) signals. Unlike traditional methods that require intricate handcrafted feature extraction, ResEMGNet takes raw EMG data as input, reducing computational complexity and enhancing practicality. Our approach was rigorously evaluated using various metrics in comparison to existing methods. ResEMGNet exhibited exceptional subject-independent performance, achieving an impressive overall three-class accuracy of 94.43\%.