Research On Cross-layer Resource Allocation For Cognitive Radio Networks

Cognitive radio (CR) is a revolution in radio technology and is viewed as an enabling technology for dynamic spectrum access. Due to the unique characteristics of the wireless networks, it is essential to address the method of multiple layers (e.g., physical, link, and network) with exploring network performance limits. The formulation of this cross-layer problem is usually complicated and challenging, while wireless resource allocation is a vital way to handle the conflict between the limited wireless resources. However, for some unique characteristics with cognitive radio networks (CRN), existing analytical technologies could not be applied directly. As a result, new theoretical results, along with new mathematical techniques, need to be developed. In this dissertation, we aim to develop some novel algorithmic design and optimization techniques that provide optimal or near-optimal solutions. The main contributions of this dissertation are summarized as follows.With the objective of maximizing data rates for a set of user communication sessions, this dissertation investigates how to design distributed algorithm for a future multi-hop CRN, The problem is developed with a cross-layer optimization approach, for joint consideration of successful transmission constraint and multi-path routing constraint. To maximize data rates for user communication sessions, the FPCR (Frequency domain-based Power Control and Routing Algorithm) is proposed, which iteratively increases the scaling factor of a session. Consequently, the design of power control, spectrum allocation, and routing modules is described. Through experimental results, the performance of the distributed optimization algorithm is compared with an upper bound and validated its efficacy.To alleviate the discrepancy between the abundance of unused spectrum in ISM bands and the overcrowding of wireless devices, the concept of a BTA (Bandwidth Time Area) is introduced to model spectrum reservation, and it is utilized to present a theoretical formalization in CRN. This concept is used to define the spectrum allocation problem as the packing of BTAs in a three dimensional time-frequency space, so that the demands of all nodes are satisfied best possible. Futhermore, the DSAT (Dynamic Spectrum Allocation with Time domain algorithm) is proposed for spectrum allocation and shows that the protocol is close to optimal in most scenarios. In terms of Virtual Round Robin, DSAT enables each node to dynamically decide on a BTA based only on local information. Using both analysis and extensive simulations, DSAT could achieve high throughput and good fragmented performance.The problem of multi-user resource management in multi-hop CRN for real time applications is addressed. A novel multi-user resource management scheme is proposed, and it allows network nodes to exchange information and that explicitly considers SUs priority scheduling. Its working principle and information swap mechanism are illustrated in detail. The proposed RDMAL (Real time Distributed Multi-Agent Learning) algorithm takes advantage of multi-agent learning approach and relies on it to dynamically exploit available band, while uses available interference information to achieve the learning efficiency. The simulation and preliminary experimental results demonstrate that the proposed RDMAL can achieve lower packet loss rate than state-of-the-art dynamic resource management algorithm and it can decrease the time overheads in terms of the cost of information exchange.For the coordination and reconfiguration in the dynamic spectrum area, a specific distributed spectrum etiquette protocol called ASCP (Adaptive Spectrum Coordination Protocol) is introduced. The reconfiguration architecture is identified for selecting from a number of potential configurations to fulfill the communication requirements of a multi-hop network. With proposed opportunistic collaboration scheme, flexible PHY/MAC and cross-layer capabilities with discovery and multi-hop routing is studied. Performance gains achieved with ASCP relative to service plane, control plane and cognitive plane in cognitive process are evaluated for example in WiFi/Bluetooth and WiFi/WiMax co-existence scenarios. Major ASCP protocol modules for bootstrapping, link artificial mapping, data path setup, naming/addressing and reconfigruration are described, and representative simulation results are provided for validation.