A Federated Deep Learning Empowered Resource Management Method to Optimize 5G and 6G Quality of Services (QoS)

Alsulami, Hemaid; Serbaya, Suhail H.; Abualsauod, Emad H.; Othman, Asem Majed; Rizwan, Ali; Jalali, Asadullah

doi:10.1155/2022/1352985

Cited by 14 publications

(4 citation statements)

References 33 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…As a new study statistic shows, the MIR of a vertex is 1400 bytes of graph payload and 3200 ns of the guard. The ns-3 discrete-event network simulator calculates Bn for each AP [35][36][37][38][39][40].…”

Section: Proposed Methodologymentioning

confidence: 99%

Aggregation-Based Dynamic Channel Bonding to Maximise the Performance of Wireless Local Area Networks (WLAN)

Parashar

Kashyap

Rizwan

et al. 2022

Wireless Communications and Mobile Computing

101

View full text Add to dashboard Cite

Channel bonding is considered by the IEEE 802.11ac amendment to improve wireless local area network (WLAN) performance. In this article, the channel bonding and aggregation method were proposed to increase wireless local area network performance (WLANs). It combines many channels (or lanes) to boost the capacity of modem traffic. Channel bonding is the combination of two neighbouring channels within a certain frequency band to increase wireless device throughput. Wi-Fi employs channel bonding, also known as Ethernet bonding. Channel bandwidth is equal to the uplink/downlink ratio multiplied by the operational capacity. A single 20 MHz channel is divided into two, four, or eight power channels. At 80 MHz, there are more main and smaller channels. Performance of short-range WLANs is determined through graph-based approach. The two-channel access techniques including channel bonding proposed for the IEEE 802.11ac amendment are analysed and contrasted. The novel channel sizing algorithm based on starvation threshold is proposed to expand the channel size to improve WLAN performance. Second-cycle throughput is estimated at 20 Mbps, much beyond the starvation threshold. Our test reveals access points (AP) 1, 2, and 4 have enough throughput. A four-AP WLAN with a 5-Mbps starvation threshold is presented. C160 = 1 since there is only one 160 MHz channel. MIR (3, 160 (a, a, a)) =0, indicating that AP 3’s predicted throughput is 0. The algorithm rejects the 160 MHz channel width since ST is larger than 0. The channel width in MHz is given by B =0,1 MIR. The MIR was intended to maximise simultaneous broadcasts in WLANs. The authors claim that aggregation with channel bonding outperforms so all WLAN APs should have a single-channel width. It usually outperforms fairness-based measures by 15% to 20%. Wi-Fi standards advise “channel bonding,” or using higher frequency channels. Later standards allow channel bonding by increasing bands and channel lengths. Wider channels enhance average WLAN AP throughput, but narrower channels reduce appetite. Finally, it is concluded that APs are more useful than STAs.

show abstract

Section: Proposed Methodologymentioning

confidence: 99%

Aggregation-Based Dynamic Channel Bonding to Maximise the Performance of Wireless Local Area Networks (WLAN)

Parashar

Kashyap

Rizwan

et al. 2022

Wireless Communications and Mobile Computing

101

View full text Add to dashboard Cite

show abstract

“…Market rule and auction theory are integrated to improve resource allocation efficiency during the evaluation. Lastly, Alsulami et al [33] introduced a federated deep learning approach to manage the resources and optimize the quality of services in 6G. Machine learning and cutting-edge technologies are widely applied for the resource allocation process in 5G, and the federated reinforcement learning approaches are incorporated with vehicle communication to improve the quality of services in 6G.…”

Section: Related Workmentioning

confidence: 99%

Improving Application Support in 6G Networks With CAPOM: Confluence-Aided Process Organization Method

Alsayaydeh,

Irianto,

Al-Gburi

et al. 2023

IEEE Access

View full text Add to dashboard Cite

Systems requiring terahertz transmission and high sampling capabilities can be supported by sixth-generation (6G) technology with minimal latency and excellent service throughput. Regardless of the distributions of data and services, High-Performance Computing (HPC) enhances speed and provides diversified applications and functionality. The Confluence-Aided Process Organization Method (CAPOM) is suggested in this article to take advantage of process allocations while using an HPC paradigm. The process allocations and completions are scheduled based on prior and current system conditions to minimize waiting time based on the 6G qualities. This implies that state requirements for process allocation, distribution, and completion are carried out with the assistance of federated learning. The initial state allocations are based on the user/application request; in other allocations, the application's request for completion time and capacity for processing are considered. Offloading and shared processing are therefore combined to maximize resource deliveries. The federated learning states are checked post-completion times to mitigate the waiting duration of dense service demands. Indicators such as distribution ratios, latency, wait time, and processing rate are considered for the effectiveness of the proofs. The suggested CAPOM achieves an 8.67% higher processing rate, 9.09% reduced latency, 8.76% less wait time, and a 6.73% higher distribution ratio for the various capacities.

show abstract

“…Figure 6 depicts an urban network that is not segregated, the coverage percentage Figure 6(a) for varied traffic densities. In densely populated areas, we may be able to achieve a high delivery rate by broadcasting data packets to virtually every automobile nearby [27,28]. Flooding causes all data packets from targeted automobiles to be retransmitted.…”

Section: An Example Of a Vanet In A Congestedmentioning

confidence: 99%

Simulation of IoT-based Vehicular Ad Hoc Networks (VANETs) for Smart Traffic Management Systems

Rizwan

Karras

Dighriri

et al. 2022

Wireless Communications and Mobile Computing

Self Cite

View full text Add to dashboard Cite

Nowadays, traffic congestion and increasing road accidents have become a major concern for both developed and developing countries. To overcome this challenge, an internet of things- (IoT-) based Vehicular Ad Hoc Network (VANET) system is proposed in which vehicles interact with other vehicles and infrastructure. These self-organized ad hoc vehicle networks not only boost traffic safety but also enhances the efficiency of traffic management systems. The VANET systems are beneficial in busy locations, where improved data dissemination protocols are used to categorize a vehicle’s transmission. However, the performance of these VANETs is also hampered by network splits and insufficient connections. Under these situations, the proposed simulation model reduces broadcast storms by minimizing redundancies, which is equally beneficial for both rural and urban settings. The suggested Next Forwarder Vehicle (NFV) protocol is based on three factors: position, distance, and orientation. Moreover, the DDP4V technique is used to analyse each of these features. Results indicate that DDP4V is compatible with 96% of all traffic scenarios and simulation durations. It can transport data packets 60% faster due to its broadcast suppression capabilities. Also, in comparison to AID and DBRS, DDP4V has fewer dispersed packets, which results into reduced retransmissions. For 200 automobiles/km2, all these techniques accomplish 35–40% of the needed cars. However, the coverage also faces some restrictions since data packets only reach vehicles on the same side as the originating vehicle. Similarly, DDP4V delivers 90% coverage for 400 vehicles/km2, which is 30% greater than prior techniques. It is observed that the proposed protocol reduces broadcast storms by employing a waggon wheel to choose the next forwarding vehicle. In high-traffic areas, it outperforms standard techniques. Similarly, it employs a patchwork of vehicles outside of the impacted region to convey data. The industrial IoT-based VANETs provide an effective tool to monitor and control traffic besides reducing the number of traffic accidents.

show abstract

A Federated Deep Learning Empowered Resource Management Method to Optimize 5G and 6G Quality of Services (QoS)

Cited by 14 publications

References 33 publications

Aggregation-Based Dynamic Channel Bonding to Maximise the Performance of Wireless Local Area Networks (WLAN)

Aggregation-Based Dynamic Channel Bonding to Maximise the Performance of Wireless Local Area Networks (WLAN)

Improving Application Support in 6G Networks With CAPOM: Confluence-Aided Process Organization Method

Simulation of IoT-based Vehicular Ad Hoc Networks (VANETs) for Smart Traffic Management Systems

Contact Info

Product

Resources

About