International Journal of Satellite-Based Communication and Wireless Networks System

Deep Reinforcement Learning for QoS and Energy Optimization in 5G Network Slicing

Minah-Eeba, Winner — 2026-07-30

Network slicing is an important feature of fifth-generation (5G) standalone systems, supporting the utilization of heterogeneous services like enhanced mobile broadband (eMBB), ultra-reliable low-latency communications (URLLC) and massive machine-type communications (mMTC) over a common setup. However, conventional static and heuristic-based slicing methods are unable to effectively adapt to dynamic traffic situations, resulting in suboptimal quality of service (QoS) and energy consumption. To address these problems, this paper suggests an AI-enabled dynamic network slicing structure based on deep reinforcement learning (DRL). A Deep Q-Network (DQN) agent is established to perform real-time resource allocation by discerning network states and optimizing a reward function that jointly reflects QoS parameters and energy efficiency. Simulation results show that the proposed method increases system throughput by up to 25%, decreases URLLC latency to below 10ms, improves packet delivery ratio to 97%, and declines overall energy consumption by approximately 12–17% when compared with conventional approaches. These findings ratify the effectiveness of reinforcement learning for adaptive and energy-efficient network control and highlight its potential to allow autonomous, scalable, and efficient resource orchestration in 5G and upcoming AI-native wireless networks.

ROMEO: Raw Onboard Multispectral Earth Observation-An Edge-AI Framework for Latency- and Bandwidth-Efficient Satellite Sensing

Charan M. H. — 2026-10-14

Whether an orbital observation proves useful often depends less on what the instrument recorded than on how long the finding took to reach somebody able to act on it. Under present practice, that interval is dominated by two serial obligations, neither involving sensing: everything the instrument produces must cross a narrow channel, and a lengthy sequence of radiometric and geometric adjustments must then be completed before interpretation may begin. This study proposes ROMEO, which discharges the interpretive obligation in flight and applies it to measurements that have received no correction. The design comprises four stages: instrument interface, a minimal conditioning layer, a divided inference engine, and a controller governing what leaves the vehicle. Division of the inference engine distinguishes the proposal — a shallow admission network, the Gatekeeper, rejects tiles showing cloud, open water, or unremarkable ground, and only survivors reach a deeper segmentation network, the Mapper. Bands are brought into register by Coarse Spatial Coregistration, which replaces runtime correspondence search with constants tabulated beforehand, finishing a granule in roughly 16.63 ms on a CPU and 1.65 ms on a GPU at sub-pixel residual displacement. Supervision derives from THRawS and SegTHRawS. Because conclusions rather than imagery are released, the quantity requiring transmission falls from gigabytes to kilobytes.

Development of a Low-Complexity Fuzzy Logic-Based Packet Scheduler for URLLC in 5G New Radio Networks

Akinlua Oluwatoyin Esther — 2026-10-21

Ultra-Reliable Low-Latency Communication (URLLC) obliges the 5G New Radio Medium Access Control (MAC) scheduler to finish resource allocation within roughly 50 microseconds (µs) of every Transmission Time Interval. Existing fuzzy schedulers reported in the literature do not measure their per-decision computational cost against this budget, and are tuned to delay thresholds an order of magnitude more relaxed than the 1 millisecond (ms) URLLC deadline. This study introduces a Low-Complexity Fuzzy Logic-Based Packet Scheduler (LCFS) built on two normalized inputs, six triangular membership functions, a nine-rule Mamdani inference engine, and binary priority classification, capped at 61 arithmetic operations per user per decision. Tested in a single-cell 5G New Radio downlink with Poisson arrivals and Rayleigh fading against Modified Earliest Deadline First (EDF), Exponential Proportional Fair (EXP/PF), and a Conventional Fuzzy Logic Scheduler (CFLS) built from the published FUSE scheme, over 12 scenarios (100 Monte Carlo trials apiece), the proposed scheduler missed the deadline for 0.0870% of packets on average (reliability 99.913%), and its delivered packets had 0.2655 ms mean delay and 0.6862 ms mean 99th-percentile delay; because delay is measured over delivered packets only, the deadline miss ratio is the main basis for comparison. Its median scheduling time per TTI, averaged over the 12 scenarios, was 35.2 µs, against 54.9 µs for EXP/PF and 66.6 µs for the CFLS. Analysis of variance with Tukey post-hoc testing showed significantly lower packet delay than Modified EDF in every scenario and no significant difference in delay from EXP/PF or the CFLS. At the heaviest load, the proposed scheduler missed about 21 deadlines per 10,000 packets, against 28 for EXP/PF and 39 for the CFLS (25.6% and 45.8% fewer), but these differences were significant only in the two heaviest-load scenarios; under lighter loads, the three schedulers performed alike. Because the two fuzzy schedulers differ in inputs, classification and ranking as well as in delay threshold, the gain over the CFLS cannot be attributed to the deadline setting alone. A sensitivity analysis found no consistent effect of the scheduling-metric weight in the scenarios examined. The simulation includes no radio errors or retransmissions, so the reliability figures describe scheduling alone; the five-nines target was not reached even under this assumption.

Integrating Dosimetric Modeling and Epidemiological Analysis to Quantify the Health Risk of Radiofrequency Exposure from Mobile Base Stations: A Population-Based Cohort Study

Belay Sitotaw Goshu — 2026-10-22

The proliferation of mobile phone base stations has raised public health concerns regarding potential adverse effects of radiofrequency electromagnetic field (RF-EMF) exposure. However, previous epidemiological studies have been limited by crude exposure assessment methods that rely on distance-based surrogates rather than individual-level dosimetric estimates. The objective of the study is to develop and validate a dosimetric model for estimating individual RF-EMF exposure from mobile base stations and to quantify associations with health outcomes in a population-based cohort. This population-based prospective cohort study included 11,529 participants from the Dutch AMIGO cohort. The study estimated RF-EMF exposure using the NISMap three-dimensional geospatial propagation model and validated it against personal dosimeter measurements in a subsample of 47 participants. Health outcomes included cancer incidence (registry-linked), nonspecific symptoms, and reproductive outcomes. The study examined associations using generalized linear models, Cox proportional hazards models, and restricted cubic splines for dose-response analysis. The dosimetric model demonstrated moderate validity (Spearman's ρ = 0.49–0.53; weighted Cohen's κ = 0.52). Nonspecific symptom scores were positively associated with modeled RF-EMF exposure (adjusted β = 0.48, 95% CI: 0.21–0.75, p = 0.001), with a nonlinear dose-response relationship (p for nonlinearity = 0.04). No significant associations with cancer incidence (HR = 1.08, 95% CI: 0.84–1.39) or reproductive outcomes (OR = 1.18, 95% CI: 0.83–1.68) were observed. The integrated dosimetric-epidemiological framework provides a refined approach to RF-EMF exposure assessment. Although no associations were found with cancer or reproductive outcomes, a modest, nonlinear association with nonspecific symptoms was observed, suggesting the need to continue investigating non-thermal effects and symptom perception mechanisms.

Cole-Cole/FDSC Co-Design of a Robust Compact Dual-Band Implantable Biosensor Antenna in Dispersive Multilayer Tissue

Settapong Malisuwan — 2026-10-23

Implantable biosensor antennas are sensitive to the dispersive electromagnetic properties of surrounding tissue and to implantation variability. This study presents an uncertainty-aware co-design framework combining a band-constrained Cole-Cole tissue model, a multilayer field-filling approximation, a dual-resonance impedance surrogate, and a frequency-dependent Smith-chart (FDSC) representation. A field-integral derivation distinguishes the weights in an arithmetic complex-permittivity mixture from electric-energy participation factors. The derivation also identifies the frozen-field, tissue-loading, and dispersion assumptions behind the resonance approximation; it does not establish an exact stored-energy model for lossy tissue. The reported computational study propagates tissue, filling-factor, depth, encapsulation, feed, and fabrication uncertainty through 12,000 fixed-seed Monte Carlo samples. Relative to nominal tuning, uncertainty-aware electrical targets increase the proportion satisfying S₁₁ ≤ −10 dB at every evaluated frequency from 36.5% to 85.8% in the MedRadio band, from 33.0% to 79.8% in the 2.45-GHz band, and from 12.2% to 68.6% jointly. Median worst in-band S₁₁ changes from −7.67 to −15.39 dB and from −7.96 to −13.59 dB, respectively. These are reused-ensemble, grid-based estimates rather than independent validation or continuous-frequency guarantees. A conductivity-and-matching-scaled SAR proxy is approximately 0.20 W/kg at its 95th percentile for 1 mW input, but is neither a proven upper bound nor a permissible-power calculation. The framework supports selection of electrical targets for subsequent full-wave and phantom investigation; exact numerical reproduction additionally requires the original fitted coefficients and computational configuration.