https://repositorio.ufu.br/handle/123456789/5464 Tue, 13 Oct 2026 19:29:34 GMT 2026-10-13T19:29:34Z https://repositorio.ufu.br/handle/123456789/50599 Title: Controle adaptativo de pesos de QoS e equidade no escalonamento de downlink 5G via aprendizado por reforço profundo Abstract: Downlink resource scheduling in 5G networks carrying mixed traffic requires balancing three conflicting objectives: aggregate throughput, fairness, and deadline compliance for latency-sensitive flows. Conventional schedulers optimize one axis at the expense of the others, while quality-of-service (QoS)-aware designs rely on weights fixed at design time that do not track varying network conditions. This work presents the DRLQoSScheduler, a downlink scheduler whose latency and reliability urgency weights and fairness target are tuned at run time by an agent trained with the deep deterministic policy gradient (DDPG) algorithm. Rather than allocating resource blocks directly, the agent applies indirect parametric control to a QoS-weighted proportional fair metric, preserving the interpretability of the allocation and avoiding the per-slot optimization required by conventional approaches. The allocation is computed in closed form, with complexity linear in the product of the number of resource blocks and the number of users. The scheduler was implemented in the Vienna 5G System Level Simulator and evaluated on a three-tier heterogeneous network across four scenarios varying user density, mobility, and the latency-sensitive traffic deadline, with 30 independent runs per configuration. Across 36 two-tailed t-tests against three native baselines, the proposed scheduler is never outperformed with statistical significance and prevails in 25, with deadline compliance favorable in all 12 comparisons of that metric. Under saturation, at eight times the reference density with unchanged topology, it is the only one of the five evaluated schedulers ranking second in all three metrics simultaneously. Fri, 25 Sep 2026 00:00:00 GMT https://repositorio.ufu.br/handle/123456789/50599 2026-09-25T00:00:00Z https://repositorio.ufu.br/handle/123456789/50470 Title: Comparação entre algoritmos de aprendizado por reforço profundo e controlador PID aplicados ao seguimento de linha Abstract: This work presents an experimental comparative evaluation of six deep reinforcement learning algorithms Deep Deterministic Policy Gradient (DDPG), Twin Delayed DDPG (TD3), Soft Actor-Critic (SAC), Proximal Policy Optimization (PPO), Advantage Actor-Critic (A2C), and Model-Based Reinforcement Learning (MBRL) and a Proportional-Integral-Derivative (PID) controller, applied to the line-following task using computer vision on a differential mobile robot. The experiments were conducted in a simulated environment, considering three scenarios of increasing geometric complexity and two operating speeds, corresponding to 50% and 75% of the robot's maximum stable velocity. The evaluation considered metrics of precision, stability, control effort, completion rate, and robustness. The results indicate that DDPG achieved the best balance between accuracy and completion rate across all evaluated scenarios, maintaining consistent performance even as speed increased. The PID controller demonstrated competitive results under stable conditions, with lower control effort, but showed sensitivity to speed increases. The remaining reinforcement learning algorithms presented significantly higher tracking errors. Overall, the results reinforce the potential of deep reinforcement learning for continuous control tasks in mobile robotics, highlighting that the most suitable method depends on the specific requirements of the application. Mon, 31 Aug 2026 00:00:00 GMT https://repositorio.ufu.br/handle/123456789/50470 2026-08-31T00:00:00Z https://repositorio.ufu.br/handle/123456789/49547 Title: Implementação e avaliação experimental de sensores para detecção de descargas parciais Abstract: Partial discharges are an important indicator of the degradation of insulation systems in medium and high voltage electrical equipment, and are widely used for diagnosis and monitoring of the operational condition of these assets. In this context, this dissertation presents the implementation and experimental evaluation of sensors for the detection of partial discharges in insulated cables, developed from commercially accessible materials and additive manufacturing techniques. The objective of this study was to develop and experimentally characterize two electromagnetic sensor architectures, called Helical Antenna Sensor (HAS) and HFCT Relative Signature Sensor (HRSS), in addition to investigating the influence of physical, geometric and constructive parameters on the sensitivity of these devices. For this, design, manufacturing, experimental calibration and laboratory tests were carried out on insulated cables containing partial discharges. The performance of the prototypes was compared to that of a commercial HFCT sensor used as a reference. The results showed that both architectures were able to detect signals associated with partial discharges and reproduce patterns compatible with the investigated phenomena. It was found that changes in the construction parameters directly influence the sensitivity and response of the sensors. Fri, 17 Jul 2026 00:00:00 GMT https://repositorio.ufu.br/handle/123456789/49547 2026-07-17T00:00:00Z https://repositorio.ufu.br/handle/123456789/49504 Title: Análise e controle de um sistema fotovoltaico conectado à rede elétrica com mitigação de harmônicas e armazenamento de energia Abstract: This work presents the analysis and control of a single-phase grid-connected photovoltaic system with integrated energy storage and harmonic distortion mitigation. The proposed system consists of a photovoltaic array connected to the direct current (DC) bus through a DC–DC Boost converter, an energy storage system based on batteries associated with a bidirectional Buck–Boost converter, and a Full-Bridge DC–AC inverter responsible for interfacing with the electrical grid. As a strategy to improve power quality, different passive filter topologies, namely L, LC, and LCL, were investigated and compared regarding their effectiveness in reducing harmonic distortions produced by the inverter. The adopted methodology was based on mathematical modeling, controller design, and computational simulation using PSIM software under different operating conditions to evaluate electrical performance, DC bus regulation, and power quality indices. The impacts of filter insertion on the injected current waveform and on the overall system performance were analyzed. The results demonstrated that all evaluated topologies contributed to reducing the Total Harmonic Distortion of current (THDi), achieving values of 3.77% for the L filter, 2.20% for the LC filter, and 1.94% for the LCL filter, highlighting the superior performance of the latter configuration. Additionally, the system achieved a power factor of up to 0.9999 and maintained the DC bus voltage regulated at 400 V, indicating satisfactory control performance and proper integration among photovoltaic generation, energy storage, and grid connection. It is concluded that the combined use of control techniques and passive filtering contributes to improving power quality and enhancing the operational performance of grid-connected photovoltaic systems. Mon, 13 Jul 2026 00:00:00 GMT https://repositorio.ufu.br/handle/123456789/49504 2026-07-13T00:00:00Z