چكيده لاتين
Cellular networks are known as the most important communication infrastructure in the world catering to a diverse range of applications with varying quality of service (QoS) demands. The advent of 5G networks aims to address these demands by offering three distinct service types tailored to different QoS requirements: high data rate applications, ultra-reliable and low latency applications, and applications necessitating massive access. Moreover, upcoming applications for next-generation cellular networks underscore the importance of meeting QoS requirements across the three key domains of data rate, reliability, and latency. Consequently, thereʹs a growing imperative to adopt a holistic approach to delivering QoS requirements concurrently in future cellular network generations.
Link Adaptation is one of the techniques designed to deal with communication channel variability. This technique tries to maximize the transmitted data rate by determining the parameters of a connection like modulation order and coding rate. Traditional link adaptation has shortcomings such as not considering the quality of service required by the applications and the lack of flexibility needed to support new applications. Also, considering the smaller cells size in the next generation, concepts such as multi-connectivity seem necessary to support the seamless communication of users with the cellular network.
The current research tries to propose an intelligent QoS-based method for multi-connectivity scenarios. More specifically, the main question raised in this research is which base stations a user should connect to and how the connection parameters of each base station should be set to meet the userʹs QoS requirements. In order to answer this question, first, a process of estimating QoS parameters is performed based on the userʹs reception status and the amount of available resources for each base station. Based on this, the members of the serving multi-connectivity cluster are selected as well as their participation in the data transfer process. Then, each base station tries to perform the link adaptation operation using the deep reinforcement learning agent in such a way that the determined participation is provided. The simulation results show that the proposed method has been able to provide the QoS of the users with a high success rate and at the same time improve the spectrum efficiency compared to state-of-the-art.
This research aims to propose an intelligent Quality of Service (QoS)-oriented approach tailored for multi-connectivity scenarios. Specifically, the primary inquiry posed within this study pertains to the optimal selection of base stations for user connectivity, alongside the determination of connection parameters for each base station to align with the userʹs QoS requisites. To address this query, an initial phase involves the estimation of QoS parameters predicated on the userʹs reception conditions and the available resource allocation across base stations. Subsequently, the members of the multi-connectivity cluster are identified, along with their respective pariticpation share in the data transmission process. Following this, each base station undertakes link adaptation operations utilizing a deep reinforcement learning agent, thereby ensuring the prescribed participation levels are maintained. Simulation results demonstrate that the proposed algorithm achieves a acceptable success rate in meeting user QoS requirements while concurrently enhancing spectrum efficiency compared with prevailing state-of-the-art techniques.