Research On Broadband OFDMA Random Access Technology

The development of wireless communication technology and the emerge of Internet of Things make the next generation wireless networks exhibit the following new characteristics:ultra-high data rate, ubiquitous coverage, massive wireless links and distributed access (cloud access). Concurrently, compelled by large energy consumption and increasingly scarce spectrum, the focus of network protocol designs is shifting from the traditional data rate augment to the enhancement of communication efficiency. Random access is a major category of medium access control technology. It will be extensively applied in various communication scenarios in next generation networks such as ultra-high data rate wireless local area networks, device-to-device (D2D) communications, vehicular ad-hoc networks and wireless sensor networks. Moreover, it will play an important role in delivering traffic data from Internet of Things and the seamless handover between cells providing dense wireless network coverage. Unfortunately, the traditional random access schemes cannot fulfil the requirements of the next generation network due to their inherent frame collision issue and the large protocol overhead in high data rate transmissions. Orthogonal Frequency Division Multiple Access (OFDM A) is a type of physical layer technology extensively applied in state of art wireless communications. It allows simultaneous data transmissions of different users on multiple orthogonal sub-channels, which substantially enhances spectrum access agility and can notably amend the performance of random access systems. In view of the advantages of OFDMA and the broad application prospect of random access, this thesis studies the combination between them.This thesis first concentrates on cross-layer design of OFDMA random access protocol. The research aims to find a scheme that can make the full use of the spectrum access agility of OFDMA technology. In this context, OFDMA technology is exploited to expand the time-domain channel contention in conventional single channel random access systems to the frequency-domain channel contention on multiple sub-carriers. In addition, the contention mechanism is designed with full consideration of network frame aggregation technology and uplink synchronization requirements. Particularly, the proposed system represents the channel access request by the busy-idle statuses of sub-carriers rather than the control frames as in the traditional schemes. Moreover, the proposed scheme enables the access point to simultaneously handle a large proportion of users’ requests. Benefited by this design, the channel contention can be resolved in the frequency domain with the finer granularity so that the frame collision probability is substantially reduced. At the same time, the channel utilization of the proposed system is further improved by aggregating a large number of uplink data transmissions. This work compares the traditional OFDM A random access protocols based on control frames and the proposed random access scheme based on subcarrier sensing. The results of the latter show the larger throughput and the shorter access delay.This thesis presents the demand bandwidth control concept and further decomposes it into two sub-issues towards the maximization of system channel utilization:the number of accessible sub-channels tuning and the sub-channelization tuning. In OFDMA random access systems, demand bandwidth, i.e. the total bandwidth of sub-channels that a node attempts to access in transmissions, can affect channel utilization significantly both in terms of channel contention intensity and the ratio of protocol overhead to data transmission time. For this influence, this thesis proposes a pair of algorithms which dynamically tune the number of accessible sub-channels and the sub-channelization respectively. The related theoretical analysis verifies the importance of demand bandwidth control. The simulation experiments demonstrate that the proposed dynamic demand bandwidth control algorithms notably outperform the traditional schemes based on static channel division and empirical adjustment of sub-channels.The system delay performance is strictly required in the next generation wireless networks. In order to fulfil the requirement, this thesis develops an analytic model for protocol designers to predict and optimize the delay performance of OFDMA multi-channel random access systems. According to the characteristics of OFDMA technology, the developed analytic model regards a random access system as a special time-slotted system, where the size of a time slot is random, and a node can access multiple sub-channels simultaneously. Owing to the generality of the adopted assumptions and the removing of the dependence on a specific protocol, the established model is applicable in most OFDMA random access systems. On the other hand, system stability is also studied in this thesis for its significant influence on system delay performance and packet loss rate. The study acquires the conditions for system stability. Moreover, it also obtains the approaches to solve and approximate the parameters related to the conditions.Eventually, this thesis extends the study of OFDMA random access system to a general communication system. By exploiting a hierarchical QoE evaluation model, the study aims to establish a QoE provisioning paradigm for the networks with heterogeneous traffics and terminals. An objective of the next generation wireless networks is to establish a system, where human-to-human, human-to-machine and machine-to-machine communications coexist. In such a system that connects all things, Quality of Experience (QoE) provisioning will become intractable because of the exhibited diversity and heterogeneity in traffics and terminals. In addition to Quality of Service (QoS), QoE can be influenced by a number of non-technological factors (e.g. user preference and behaviors). Consequently, its provisioning is more difficult than that of QoS. In the proposed hierarchical model, QoE is classified into application level, node level and network level. The model adopts the satisfaction metric to evaluate the gap between the current QoE and the desire of a user with allowing the heterogeneity among traffic demands and user preferences. For validation, the proposed model is applied in a random access network with traffic and node heterogeneity to achieve the QoE-oriented optimization of the internal traffic flow scheduling and external channel contention coordination. Theoretical analysis and simulation results indicate that the proposed model can comprehensively characterize user experience from microcosmic view (application level) to macrocosmic view (network level). Moreover, the developed optimization algorithms are also proved to be effective for the maximization of user satisfactions with weighted proportional fairness.