Investigation on adaptive learning video streaming techniques to achieve good quality of experience for heterogeneous wireless network

Abstract

In recent years, real-time video streaming has gained much popularity. To cope with the progressive development of Internet of Things (IoT) and other heterogeneous wireless networks, it becomes necessary to fairly allocate the network resources among different kinds of users and to realize best Quality of Experience (QoE) and performance objectives. So far, most of the researchers have worked on Forward Error Correction (FEC) techniques to make a balance between QoE and performance. However, as network capacity increases, the performance tends to degrade affecting the live visual experience. newlinePresently, Machine Learning (ML) and Deep Learning (DL) algorithms have been successfully incorporated in FEC methods to stream videos in a perfect manner across multiple heterogeneous networks. Even so, there is scope for the improvement of the algorithms employed in video streaming system to provide better video experience without packet loss and in time. newlineIn the present research two learning models namely Extreme Learning Machine (ELM)-Adaptive FEC and Gated Recurrent Unit (GRU)-Adaptive FEC have been designed and tested. The investigation has been carried out in two phases. In the first phase, the role of ML and DL methods in FEC for perfect video streaming through heterogeneous wireless network has been studied. newline