G communication plays a crucial role in enabling high-speed, low-latency data transfer for underwater operations. Underwater communications need optimization support to overcome various problems such as packet loss and latency. This study presents a simulation model for underwater 6 G networks, focusing on the optimized placement of sensors, AUVs, and hubs. The network architecture consists of fixed hub stations, mobile autonomous underwater vehicles (AUVs), and numerous sensor nodes. Environmental parameters such as temperature, salinity, and conductivity are considered in the transmission of electromagnetic signals; signal attenuation and transmission delays are calculated based on physical models. The optimization process begins with K-Means clustering, followed by sequential application of Genetic Algorithm (GA) and Particle Swarm Optimization (PSO) to refine the cluster configurations. The simulation includes key network dynamics such as multi-hop data transmission, cluster leader selection, queue management, and traffic load balancing. To compare performance, two distinct scenarios-one with cluster leaders and one without-are modeled and visualized through a PyQt5-based real-time graphical interface. The results demonstrate that 6 G network architectures in underwater environments can be effectively modeled and optimized by incorporating environmental conditions.

The rollout of 6G networks introduces unprecedented demands for autonomy, reliability, and scalability. However, the transmission of sensitive telemetry data to central servers raises concerns about privacy and bandwidth. To address this, we propose a federated edge learning framework for predictive maintenance in 6G small cell networks. The system adopts a Knowledge Defined Networking (KDN) architecture in Data, Knowledge, and Control Planes to support decentralized intelligence, telemetry-driven training, and coordinated policy enforcement. In the proposed model, each base station independently trains a failure prediction model using local telemetry metrics, including SINR, jitter, delay, and transport block size, without sharing raw data. A threshold-based multi-label encoding scheme enables the detection of concurrent fault conditions. We then conduct a comparative analysis of centralized and federated training strategies to evaluate their performance in this context. A realistic simulation environment is implemented using the ns-3 mmWave module, incorporating hybrid user placement and base station fault injection across various deployment scenarios. The learning pipeline is orchestrated via the Flower framework, and model aggregation is performed using the Federated Averaging (FedAvg) algorithm. Experimental results demonstrate that the federated model achieves performance comparable to centralized training in terms of accuracy and per-label precision, while preserving privacy and reducing communication overhead.

Aeronautical Ad-hoc Networks (AANET) is a fairly new concept that connects airplanes via wireless air-to-air links, allowing passengers to access the Internet during a flight. The unstable air-to-air link characteristics and ultra-dynamic topology become the main differences between AANETs and usual ad-hoc architectures. To handle these differences, the AANET topology could be created in the form of clusters by collecting airplanes having similar features under the same set. However, it is also difficult to sustain the cluster topologies since ultra-dynamic characteristics still affect them. Therefore, the current cluster topology must be continuously mapped to the airplanes to notify them as a part of sustainability. If we do not ensure the sustainability of the clusters, the packet transfer success of AANET is decreased with higher end-to-end latency. At that point, to solve this aircraft notification problem and map the current cluster topology to them at each timestamp, in this paper, we propose a Gated Recurrent Unit (GRU)-based topology mapping mechanism for AANETs. Here, the GRU can continuously notify the airplanes at each timestamp about topology changes. Therefore, airplanes can forget the old topology when it changes. Otherwise, the topology taken from the previous timestamp is continuously remembered if it does not change. Finally, the performance evaluations reveal that the GRU-aided topology mapping can achieve roughly 42% higher packet delivery ratio with 34% reduced end-to-end latency.

Gamification is one of the new trend in software development and it has already gained a well-deserved popularity in finance, healthcare, education and even manufacturing. Software testing is a continuous cycle layered with several stages and spanning across multiple types of testing. Teams need to design test suites and implement test execution methodologies in each stage of development. For this reason, software testing teams comprise many individuals skilled in different aspects of software testing. The inclusion of gamification in this course can lead to positive benefits based on the idea that it is used to influence behavior. This paper presents an preliminary study of Machine Learning (ML) approach for predicting performance gamification of software tester specialists under a gamified testing environment. ImonaGame is a software company that delivers gamification as a service for software testing team of 30 members with different static and dynamic data. User behavior collected in dynamic data sets was classified into categories by deconstructing complex activities into behavior chains using supervision of domain experts. The classification approach was centered on the system’s testing processes’ performance objectives and potential for encouragement or dissuasion. Motivators and obstacles for the target activity and its behaviors will be found when the model has been developed. After conducting preliminary research, it is possible to determine whether gameful design is an effective and efficient tactic for achieving the desired result by analyzing needs, motives, and obstacles. The source data was classified target data in four categories such as I: Static feature (personal information; 4), II: Daily feature (gamification elements; 14), III: Mission feature (points; 7 sources) and IV: Cumulative futures (Sum of daily and mission features; 13).