HYBRID TCP RENO–VEGAS CONGESTION CONTROL FOR RELIABLE DATA TRANSMISSION IN IOFT-BASED UAV NETWORKS

Abstract

Under Network-Centric Warfare (NCW) paradigm, low-latency, high reliability and reconfigurable communications is required to facilitate persistent information exchange between geographically dispersed aerial and terrestrial communication nodes. Because of the role of the Unmanned Aerial Vehicles (UAV) that enable the existence of Unmanned Aerial Vehicle assisted communication networks, The Internet of Flying Things (IoFT) system has experience extremely fast-varying network topology where topology can changes very rapidly, there can be an instantaneous network disconnection and wireless channels are in a highly dynamic and random manner. The conventional Transmission Control Protocol (TCP) based congestion control algorithms (CCAs) designed for wired networks fail to be performed adequately under such scenarios. The basic TCP Reno algorithm is a loss based CCA that performs efficiently with packet loss recovery but treats the packet loss occurring on wireless link due to transmission impairments as the cause of congestion and so exhibits unstablie throughout. On the other hand, the TCP Vegas algorithm applies delay based estimation of congestion which leads to the stability in the rate and thus may cause less efficient utilization of bandwidth under highly dynamic environment. In this work, a Hybrid TCP Reno-Vegas congestion control model is proposed, that combines the proactive delay based estimation of TCP Vegas and effective loss recovery of TCP Reno so that to enhance transport-layer communication over IoFT-based UAV network. The proposed model was simulated on network simulator-2 (NS-2) under same mobility model, routing and traffic scenario in the context of Unmanned Aerial Vehicle assisted Mobile Ad Hoc Network (MANET).Performance evaluation was conducted using throughput, congestion stability, packet delivery behavior, and transmission efficiency metrics. Simulation shows the hybrid scheme proposed has the best throughput behavior and the stable transmission over TCP Reno and TCP Vegas. With a hybrid model, an average throughput close to 2.28 Mbps is gained, which exceeds 2.02 Mbps for TCP Reno and 1.78 Mbps for TCP Vegas, and a better congestion window adaptation and bandwidth utilization is achieved. These findings indicate that the proposed Hybrid TCP Reno–Vegas mechanism provides an effective congestion control solution for reliable and delay-sensitive communication in IoFT-enabled NCW environments.

Keywords:

Internet of Flying Things (IoFT); Network-Centric Warfare (NCW); Hybrid TCP Reno–Vegas; Congestion Control; UAV Networks; Mobile Ad Hoc Networks (MANET); Throughput Optimization; Transport Layer Performance; NS-2 Simulation