Adaptive Video Task Offloading Based on Mobile Vehicle Assistance

With the increasing number of Internet of Things(IoT) devices, many computationally intensive vedio tasks need to be processed. Mobile edge computing(MEC) has emerged as an effective solution, but a large number of real-time video tasks are still in the queue especially in congested traffic areas or busy periods. Considering the abundance of idle computing resources in moving vehicles on roads, this paper utilizes moving vehicles to assist edge servers in offloading video tasks. Firstly, the dwell time of vehicles calculated by the movement trajectory of vehicles is used to select candidate vehicles for task offloading. Secondly, the channel quality for video task transmission is analyzed and network bandwidth is predicted. Additionally, a QoE model is constructed, and we propose a bitrate adaptive selection algorithm that can transmit videos with the highest possible resolution. The simulation proves that our proposed method can provide more efficient and stable offloading services.

I. Introduction

The continuous development of 5G communication has en-abled higher data rates and reduced latencies, popularizing in-teractive applications like 3D video streaming, crowd-sourced streaming, AR, and VR. At the same time, the deployment of Internet of Things (loT) devices is also increasing, with loT devices being widely used in scenarios such as smart monitoring, smart cities, and autonomous driving. It is reported that by 2025, the data generated by loT devices will increase by 28% [1]. These devices will generate a large number of tasks, especially in busy traffic areas or during congested times. A significant portion of these tasks is real-time video stream processing tasks, which require substantial computing resources. Real-time video tasks typically require low latency, high reliability, and a large amount of computing resources.