The Third Generation Partnership Project (3GPP) introduced Cellular Vehicle-to-Everything (C-V2X) for vehicular communications. In the standard, C-V2X Mode 4 is defined for the distributed resource selection. Subsequently, in 3GPP Release 16, NR-V2X is introduced with Mode 1 and Mode 2 for vehicular communications. Likewise C-V2X Mode 4, NR-V2X Mode 2 is used for decentralized resource scheduling. The vehicles select the resources based on their local observations by utilizing the Semi-persistent Scheduling (SPS). Since, the vehicles select the resources based on the local observation, sensing nature of SPS is challenged by the hidden node problem that lead to resource conflict. To resolve the contention, 3GPP also introduced the Physical Sidelink Feedback Channel (PSFCH) to assist the distributive resource scheduling based on the receiver feedback. However, this incurred a signaling overhead. In this work, federated learning is exploited for distributive training via offline method and distributive multi-agent-based resource scheduling is performed following the principles of NR-V2X Mode 2. Distributed training favors the model accuracy by accommodating the varying affect of the environment due to the high mobile dynamics. Simulation is conducted by integrating SUMO in conjunction with 3GPP NR-V2X standard. Performance results demonstrate a substantial improvement compared to other deep learning methods, where centralized training and random resource selection procedures are employed. This research marks a significant stride towards efficient and conflict-resilient resource allocation in vehicular communications.

5G revolutionizes connectivity with network slicing, but non-uniform slice deployment poses challenges. Efficient slice handover mechanisms and predictive learning offer solutions. We propose Long Short-Term Memory (LSTM) based prediction of network slice traffic demands for proactive resource management to resolve challenges of mobility management and non-uniform slice deployment within a registration area. Thereby, allowing seamless service provision throughout the registration area. Our proposed model shows excellent results in predicting the number of received requests for all available slices. These results are then utilized to deduce the average slice demand for each service in the network. The obtained results are beneficial for configuring the slice resources proactively in order to provide seamless service to mobile users throughout the registration area, regardless of the non-uniformity in slice deployment among different tracking areas.

The third-generation partnership project (3GPP) introduces the 5G NR-V2X to supplement the C-V2X to support advanced applications. 3GPP defines the semi-persistent scheduling for distributed resource scheduling likewise in C-V2X, however new medium access control (MAC) features are introduced in NR-V2X mode 2. Re-evaluation mechanism is added in semi-persistent scheduling to reduce resource contention. Despite with new MAC feature, NR-V2X mode 2 cannot handle the scheduling of aperiodic packets efficiently. With the increase in vehicular density, channel congestion occurs leading to packet collision. 3GPP defines the channel congestion control mechanism based on two metrics; channel busy ratio (CBR) and channel occupancy ratio (CR). These metrics, however, have considered the system-level requirements but ignore the application-level requirements such as age-of-information (AoI) associated with the message packet. In this paper, we proposed a deep reinforcement learning-based congestion control mechanism to support both system and application requirements. The performance of the proposed scheme is evaluated and compared with the conventional decentralized congestion control mechanism in a simulator designed inline with the 3GPP specifications.

Scalability presents a formidable challenge in traditional Cell-Free (CF) massive Multiple Input Multiple Output (mMIMO) networks, driven by escalating computational demands on access points (APs) and the reliance on a single central processing unit (CPU). To address this, the study proposes a dynamic cooperative clustering (DCC) method, tailored for both backhaul (CPUs-APs) and fronthaul (APs-Users). In the backhaul phase, DCC strategically pairs APs with CPUs using the Kuhn-Munkres algorithm, ensuring equitable resource allocation by considering distance matrices, channel statistics, APs traffic load, and available CPU resources, thereby fairly balancing the distribution of computational load across the CPUs. Subsequently, in the fronthaul phase, the focus is on optimizing the selection of APs for user-centric clusters, using Particle Swarm Optimization (PSO). This optimization aims to maximize the overall sum rate while intelligently managing the inclusion and exclusion of APs within each user-serving cluster. Through extensive simulations, the study highlights the potential of the proposed approach to address scalability concerns in CF-massive MIMO systems, promising improved performance in wireless communication networks. The comparative analysis demonstrates the superiority of the proposed scheme over conventional clustering schemes, consistently delivering better sum rates across various scenarios, with an 18.23% improvement in sum rate and a 30% enhancement in Load Balancing Index (LBI), indicating significantly improved resource distribution and network efficiency.

Vehicle ad hoc networks (VANETs) is an emerging ad hoc network technology with a promising future but significant limitations, particularly in communication networks. An effective routing system can significantly enhance the performance of VANETs. However, creating an effective routing protocol in an urban context to transport the data packet to the destination is still challenging due to complex road conditions, sporadic connectivity among vehicles, frequent disconnections, and quick changes in network topology. To enhance the accuracy of routing decisions in an urban setting, we propose a novel “multiple attributes decision-making junction selection routing (MADMJSR)” routing protocol based on vehicle position for VANETs. Formerly, most of the research was focused on static weight-based attribute selection for communication rather than optimal values, which led to a drop in junction decision-making accuracy. Our proposed scheme considers the multiple attribute selection and the optimal weight value assigned to each attribute, which significantly improves the routing decision accuracy. The results show a high packet delivery ratio (PDR) gain within minimum latency and less network load. Furthermore, a thorough discussion of the functioning of our proposed protocol is presented along with a detailed performance evaluation in comparison to other protocols.