Font Size: a A A

Research On Dynamic Spectrum Access Models And Methods In Cognitive Radio Networks

Posted on:2011-07-10Degree:DoctorType:Dissertation
Country:ChinaCandidate:Y LiFull Text:PDF
GTID:1118330368488045Subject:Signal and Information Processing
Abstract/Summary:PDF Full Text Request
In recent years, with rapid development of wireless communication technology, wireless spectrum is increasingly becoming a scarce resource. When many wireless access technologies (Wireless Local Area Networks (WLAN), Wireless Personal Area Networks (WPAN), Wireless Metropolitan Area Networks (WMAN)) put into application, users set a still higher demand on wireless frequency resources. At present, the radio spectrum is managed by government, and a fixed allocation mode is adopted. Radio spectrums are divided into licensed frequency and unlicensed frequency. Due to rapid development of various wireless technologies which using unlicensed frequency, the unlicensed frequency is getting more congested. On the contrary, in many regions and time periods, many licensed bands have not been fully used by their users (spectrum holes). Dynamic Spectrum Access (DSA), which provides wireless service by using spectrum holes, is put forward to handle the contradiction. Cognitive Radio (CR) is a key technology that enables the wireless network which adopts DSA technology to capture spectrum holes.CR network in this dissertation is limited to the wireless network which provides wireless service using spectrum holes. CR networks share wireless spectrum with licensed user networks and each has different priority, which cause time-varying property of spectrum resource and network topology. Therefore, it is necessary to manage the shared spectrum efficiently in CR networks.This dissertation studies the dynamic topology model and time-varying link model of CR networks. Moreover, spectrum sensing is the basis of sharing spectrum. How to improve sensing efficiency is another research emphasis. The main contents of this dissertation are organized as follows.(1) We introduce the developments, characteristics and application scenarios of DSA and CR in this dissertation. Their research developments in detail as respect of spectrum sensing, spectrum management, spectrum handover, spectrum sharing, and upper layer issues are reviewed, and their research difficulties as well as their research directions are also analyzed.(2) For the dynamic property of available spectrum in CR networks, the topology of CR networks is also time-varying. Dynamic Layered-Graph (DLG) model is proposed to describe the dynamic topology of CR networks. In DLG, layer is corresponding to channel, the state of sub-node on a layer is corresponding to the avalibilty of the channel at node which the sub-node belongs to. The state of sub-node is transferred according to a time-varying link model. Simulation results show that the DLG model can describe CR network accurately and improve routing success rate.(3) On the basis of the DLG model, a cross-layer design based routing algorithm is studied. It is an integrated design of spectrum decision, interface allocation, route finding and DLG updating. With the aid of interface allocation algorithm and dynamically updating the weights of edges in DLG, the routing algorithm do it best to select the node with idle interface and channel with lower load to improve the ability of full-duplex communication and load balance. If a route must share interface with others, it should reduce the interface switching among channels.(4) The variation law of the availability of link in CR networks is deeply studied. A time-varying link model based on nested semi-markov link and markov link is proposed. An improved LZ78 algrithm is developed to learn the statistical law of licensed users occupying channel duration and spectrum holes duration, and to update model parameter according to new coming data.(5) Based on time-varying link model mentioned above, a simple yet efficient algorithm is development to provide reliable short-term and long-term spectrum usage prediction was studied. Simulation results show that the proposed algorithm can run well under the different distribution of the duration of spectrum hole.(6) It is important to allocate a reasonable spectrum hole to a CR user with QoS requirement. A QoS provisioning spectrum decision algorithm is proposed. Rather than presuppose the distribution of the duration of spectrum hole, the proposed algorithm learn the distribution with aid of time-varying link mode. Base on the spectrum usage prediction algorithm and take the request time of a user and the age of spectrum hole at the time of the user initiating the session into accounts, to guarantee Qos in a sense of spectrum handover probability.(7) An adaptive spectrum sensing scheme is proposed. Sensing time was divided into many time-slots. After a time-slot sensing, the action set consists of data transmission, spectrum handover and continue sensing. The action is selected in the action set by comparing the sensing result with corresponding elements of two threshold vectors. Simulation results show that the proposed algorithm achieves a higher spectrum utilization compared to spectrum sensing scheme with fixed sensing time.
Keywords/Search Tags:Cognitive Radio, Dynamic Spectrum Access, Network Topology Model, Time-varying Link Model, Spectrum Usage Prediction, Spectrum Decision, Quality of Service, Adaptive Spectrum Sensing
PDF Full Text Request
Related items