Research On Quality Of Experiences Based Wireless Resource Management

The advent of the mobile internet has sparked an ever increasing interest in mobile wireless multimedia applications. Due to the scarcity of wireless network resources, the wireless channel interference and the transmission delay, how to utilize the radio resource effectively and provide heterogeneous services to users over time-varying radio channel is a challenge to the system design. The radio resource management which supports heterogeneous services with different QoS (Quality of Service) requirements plays a key role in addressing this challenge.This thesis studies the QoE (Quality of Experience) based wireless resource allocation in the wireless communication networks. We propose solutions for the following aspects:how to evaluate the QoE properly, how to design the wireless resource allocations to improve the QoE, and how to improve the energy efficiency with the QoS guarantee through the wireless resource allocation.Firstly, the evaluation of the user’s experience is the basis of the QoS based radio resource allocation management. The imprecision and immeasurability of the user’s feeling are the main challenges that the search for proper expression of a user’s satisfaction shall meet. In this thesis, we leverage on the application of fuzzy control theory to quantify the users’satisfaction from a collective set of fuzzified QoS parameters, which we called fuzzy satisfaction factor (FSF). The proposed FSF is a combination of outputs from several human feeling rules which allows to take into account the uncertain observations, and carries out a comprehensive decision with high performance.After that, radio resource management is an important means and key technology to improve the QoE. Therefore, to improve QoE, the radio resource management which supports heterogeneous services with different QoS in single cell and multi-cell environments are carried respectively.In the single cell radio resource management, the effective capacity of the multi-user OFDMA (Orthogonal Frequency-Division Multiple Access) system for time-varying channel is characterized. Then to maximize the system average effective capacity, through the Tayler approximation, an effective capacity (EC) asymptotical optimal user scheduling is proposed, which selects users based on the relationships among the moving speed, traffic profile and the effective capacity. It is further shown that the proposed scheme can improve the effective capacity of the mobile hosts. Moreover, to reduce the system complexity, by considering the property of the utility variation, we proposed the Maximize Utility Gradient algorithm (MG), which with much reduced complexity approaches the optimal capacity, while significantly improves the user experience in the time-varying channels.Multi-cell environment will bring the handovers between cells. If users suddenly switch to the neighbor cells, as the multi-cell resource allocation didn’t reserve resource to the users, it will increase the interruption duration, and significantly deteriorate the mobile users’experience. To address this problem, we present a new mobile resource reservation protocol (MRSVP) based multi-cell resource allocation scheme for OFDMA systems. Mobility model, service QoS constraints, inter-cell interference and channel state information are jointly integrated into the proposed cross-layer design framework to dynamically reserve resource, perform radio resource allocation for multiple users. This framework is capable of offering guaranteed services, avoiding multi-cell interference, while maximizing the system throughput.Previous research focused on the QoS guarantee, while with the demand for energy saving, maximizing the energy efficiency (EE) while satisfying certain QoS requirements is also desirable. By providing us the performance in the presence of soft QoS constraints, EC enables us to find the relationship between the EE and the QoS constraints. Effective capacity (EC) is defined as the maximum constant arrival rate that a given time varying service process can support while providing statistical QoS guarantees. Based on EC and taking a realistic base station power consumption model into account, we develop a novel EE metric:effective energy efficiency (EEE), to represent the delivered service bit per energy consumption at the upper layer with QoS constraints. The EEE and EC tradeoff is discussed and the effects of diverse QoS parameters on EEE are investigated. After that, maximizing the EEE problem with EC constraints for single user environment is addressed and then an optimal power control scheme is proposed to solve it. Then we further extend to the multiuser scenario to propose a cross layer resource allocation scheme, which maximizes the EEE with resource constraints. This work gives an integral resource allocation scheme which can achieve a high EE while providing a good service in each QoS aspect. Based on the QoE evaluation, this thesis studies the radio resource management from single cell to multi-cell environments. Furthermore, the relationship between QoS and energy efficiency is studied. All the researches in the thesis will help comprehension of resource allocation with QoS guarantee and cast some bright light on the design of the practical systems.