Uma proposta de algoritmo de escalonamento para o tráfego uplink em redes LTE

• Autor(a) principal: Araújo, Fabrício Geraldo
• Data de Publicação: 2014
• Tipo de documento: Dissertação
• Idioma: por
• Título da fonte: Repositório Institucional da UFU
• Texto Completo: https://repositorio.ufu.br/handle/123456789/14573; https://doi.org/10.14393/ufu.di.2014.38
• topic: LTE; Uplink; Modelagem; Simulação; LTE-Sim; Algoritmo de Escalonamento; Modeling; Simulation; Scheduling Algorithm; CNPQ::ENGENHARIAS::ENGENHARIA ELETRICA

The Long Term Evolution (LTE) network standardized by the 3rd Generation Partnership Project (3GPP) has as main objectives: achieving a high throughput, improved spectrum efficiency and reduced latency. For this, there was an evolution of the radio interface and network architecture compared to previous mobile cellular networks. Based on 3GPP s Release 8, LTE uses the multiple access technique Single Carrier Frequency Division Multiple Access (SC-FDMA) in uplink traffic, thus, resources allocated to a single user must be contiguous in the frequency domain. The efficient resource allocation for LTE network is a challenge and requires full attention from the scheduler. The main scheduling algorithms found in the literature use a matrix constructed in the time domain based on two paradigms, Channel Dependent (CD) and Proportional Fairness (PF), for the filtering of users that will be scheduled later. But, both do not consider the Quality of Service (QoS) requirements in their approaches. In view of this, we propose a scheduling algorithm to meet these requirements. The proposed mechanism considers the Modulation and Coding Scheme (MCS) as well as aspects about the types of services during the construction of the matrix. Thereafter, in the frequency domain, the resource allocation is based on the Riding Peaks algorithm. Finally, the proposed solution was evaluated using modeling and simulation. The simulation results show that the proposed scheme achieves improved performance for real-time applications when compared to other algorithms found in the literature.