Resource Allocation in Heterogeneous Networks

Student thesis: Doctoral ThesisDoctor of Philosophy (PhD)

Abstract

Due to the heterogeneity and new capacity demands in fifth generation (5G) networks, a heterogeneous network (HetNet) is being considered by the long-term evolution (LTE)-advanced to satisfy these network requirements. Due to a mix of high and low power base stations, spectral and energy efficiencies could be improved as compared with the traditional cellular network. Whereas the HetNet seems to be an efficient approach, there are still research questions on how resource allocation in such a network could be enhanced. Another challenge is the usage of unlicensed bands in future wireless networks. The coexistence between the LTE and WiFi networks within the unlicensed spectrum should be considered.
In this thesis, a system-level (SL) simulator has been developed based on the Monte Carlo simulation approach to model downlink LTE cellular networks. Key user throughput statistics such as peak, mean and edge throughputs are used for the performance evaluation throughout this thesis. The small cell deployment is first investigated by this developed simulator. By increasing the number of femtocells, the peak throughput could be around eight times as large as the conventional cellular network. User throughputs can be gained by the resource block (RB) scheduling as well. The best-channel quality indicator (CQI) scheduler appears to be the best choice to boost mean and peak throughputs as compared to the round-robin (RR) scheduler.
The simulator is further developed to model the carrier aggregation (CA). Component carrier (CC) selection and RB scheduling are considered separately. Because of the different load balance across CCs, the peak throughput of the RR CC selector could be around three times as large as the reference signal received power (RSRP) CC selector. The combination of frequency bands and RB scheduling in the CA deployment are also investigated.
Another focus of attention is the usage of unlicensed spectrum in the LTE network. It is called the unlicensed-LTE (U-LTE) in this thesis. The frame-based listen-before-talk (LBT) algorithm is included in the developed simulator to coexist the LTE and WiFi networks in the unlicensed spectrum. The novel resource allocation for the U-LTE is proposed. The LTE transmission probability on CCs derived from the WiFi network is defined for the clear channel assessment (CCA) and CC selection. The transmission probability threshold is used instead of the energy detection (ED) threshold. Subsequently, the proposed CC selectors, including maximum probability (MaxProb), random and RR, are examined. Due to the best CC load balance, the RR CC selector could offer around a twofold increase in mean and peak throughputs of the MaxProb CC selector. To investigate the impact of RB scheduling on the performance metrics in the U-LTE, user throughputs, including minimum, mean and maximum, between the macro and femto networks are compared. For the femto network, the highest user throughput could be obtained by the best-CQI scheduler.
In summary, key user throughput statistics can be significantly enhanced by the small cell deployment and CA as mentioned above. Load balancing across CCs is preferable to offer higher user throughputs. It is found that the choice of RB scheduling should be considered in connection with user throughput, fairness and computational complexity. In addition to CC selection and RB scheduling, there is still room for improvement in the coexistence between the LTE and WiFi networks within the unlicensed spectrum for both modelling and real implementation.
Date of Award23 Jun 2020
Original languageEnglish
Awarding Institution
  • The University of Bristol
SponsorsRoyal Thai Government
SupervisorAngela Doufexi (Supervisor) & Simon M D Armour (Supervisor)

Keywords

  • Heterogeneous Networks
  • LTE-A
  • Resource allocation
  • System-level simulation

Cite this

Resource Allocation in Heterogeneous Networks
Jinaporn, N. (Author). 23 Jun 2020

Student thesis: Doctoral ThesisDoctor of Philosophy (PhD)