24 June 2020 in Journal of Optical Communications and Networking
Hybrid light fidelity (LiFi) and wireless fidelity (WiFi) networks provide a promising solution for future indoor wireless communication. Hybrid networks face the challenge of imbalanced cell load due to the uneven distribution of user equipment (UE), and they need to suppress inter-cell interference induced by the dense deployment of LiFi access points (APs). In this paper, with the aim to efficiently establish UE-AP associations and allocate resources, we propose a potential game (PG) aided design of joint user-centric LiFi coordinated multipoint (CoMP) joint transmission (JT) clustering and load balancing (LB). In our design, we formulate the joint design problem as a mixed-integer nonlinear programming (MINLP) problem, which integrates the requirement for the minimum data rate of each UE. With the aim to maximize the system throughput under proportional fairness constraints, we propose to build a PG model to solve the MINLP problem. Then we derive the Nash equilibrium based on the best-response dynamics, and thus a suboptimal solution is attained. Thanks to the space diversity gain brought by our joint LiFi CoMP-JT clustering and LB design, the signal-to-interference-plus-noise ratio can be improved. Thus, when the physical layer configurations remain the same, hybrid networks can achieve larger capacity and better reliability performances. Simulation results validate that hybrid networks using our presented design would achieve better throughput and satisfaction performances than benchmarks.
10 September 2018 in IEEE Communications Letters
The hybrid light fidelity (LiFi) and wireless fidelity (WiFi) networks is regarded as a promising solution for future indoor wireless communication. In this letter, we investigate the traffic load balancing (LB) issue for this hybrid networks. In our design, considering that user terminals (UTs) may have different data rate demands, we utilize the complementary advantages of the wide bandwidth of the LiFi network and the wide coverage area of the WiFi network to serves the UTs. With considerations of the access point (AP) heterogeneity and UTs’ quality of experience (QoE) demands for data rate, we then formulate the LB issue as a mixed-integer non-linear programming (MINLP) problem with the aim of maximizing the system throughput under proportional fairness constraints, and propose a distributed optimal algorithm and a simplified suboptimal sequential algorithm to solve it. Simulation results demonstrate that using the presented QoE-driving LB scheme, more UTs are served with higher QoE while maintaining high system throughput and proportional fairness.
25 May 2021 in IEEE Transactions on Mobile Computing
Downlink mobile edge computing (MEC) networks are requiblack to serve increasing large number of Internet of Things (IoT) devices with limited battery capacity. In order to serve massive user equipments with low power consumption requirements, in this paper, we propose an energy-efficient multi-carrier non-orthogonal multiple access (MC-NOMA) design which allows more than two IoT devices to multiplex and access the same subcarrier band. With the aim to minimize the total transmit energy while meeting the demands of each IoT device such as the low latency, in our design, we first derive the optimal successive interference cancellation (SIC) policy and minimum power allocated to every IoT device. Then we propose an optimal greedy algorithm to allocate the frequency blocks, and formulate the optimization of the computational resource allocation as a min-max problem. Subsequently, we characterize the MC-NOMA network with the potential game model, and present a scheduling scheme to manage massive IoT devices. Simulation results demonstrate that our proposed scheme can consume 3-10 dB less energy in a MEC network deployed with 256 IoT devices compablack with the conventional orthogonal multiple access (OMA) scheme and non-orthogonal multiple access (NOMA) scheme.