Design and development of MAC and routing protocols for cognitive radio based wireless mesh networks
How, Kiam Cheng
Date of Issue2014
School of Electrical and Electronic Engineering
Wireless Mesh Networks (WMNs) are communications networks made up of radio nodes organized in a mesh topology. It is envisaged that WMNs can provide broadband wireless access and replace traditional wired based communications infrastructure in the near future. Correspondingly, WMNs are also expected to support delay sensitive and/or high bandwidth real-time streaming multimedia applications such as VoIP and video conferencing. However, Quality-of-Service (QoS) provisioning in WMNs is very challenging due to various factors. These include physical layer factors such as fading or attenuation of a transmitted signal over the wireless medium which results in a loss of signal power at the receiver. Other factors include the dynamical changing of topology, capacity limitations, link variability and multi-hop communications. On the network layer, many existing dynamic routing protocols select paths based on only one criteria (shortest hops, delay etc), which do not take into account topology, interference or traffic pattern. This causes the traffic to flow through nodes that experience high interference and thus increases the occurrences of dropped packets and retransmissions. Despite on-going research efforts, one key limitation to the performance of the WMNs is the interfering nature of wireless transmissions which degrades the network capacity. This is made worse, especially for the case of devices operating in the unlicensed spectrum where it is becoming increasingly crowded due to the proliferation of diverse wireless communication systems. The overcrowding of the unlicensed spectrum aggravates the interference issues which further reduces the performances of wireless devices. The interference is further exacerbated in WMNs where traffic takes multiple hops to reach the destination. Based on surveys that have shown that vast portions of the licensed spectrum remain underutilized across frequency, space and time, the use of cognitive radio (CR) based WMNs or cognitive radio networks (CRNs) is becoming increasingly popular due to their capability to exploit frequency bands currently unoccupied for use. Notwithstanding the great potential, there are also pitfalls in the adoption of CR technology. In the course of our research, we have identified many gaps in the area of WMN and CRN research literature. There were inadequacy in terms of medium access control (MAC) and network layer solutions for QoS provisioning in both WMNs and CRNs. Furthermore, many schemes proposed for CRNs have introduced significant security loopholes and this has introduced significant challenges to the operation of the CRN across various network layers. To address these gaps, we examined and proposed different strategies to tackle the challenging issue of QoS provisioning in CRNs across the three different networking layers, physical, MAC and network layer. In our work documented within this thesis, we have improved the capability of CRNs to meet the required QoS for broadband applications by designing and developing MAC and network layer protocols that provides superior performance compared to existing protocols. Our protocol design considers a combination of different metrics like network topology, traffic pattern, interference and transmission properties like transmission power, modulation, rate etc., instead of traditional shortest path approach. Our first task was to design and implement the MAC-Layer QoS Provisioning Protocol (MQPP) for Cognitive Radio Networks which combines adaptive modulation and coding with dynamic spectrum access. The simulation results have indicated that the proposed MQPP scheme can provide a significant performance improvement in terms of lower delay and higher throughput at the cost of only a slight increase of transmission power and protocol overheads. MQPP can also provide clear service differentiations for different traffic classes under all traffic loads. Next, we examined resource allocation strategy by proposing an Elastic Bandwidth Allocation Scheme (EBAS) with a consideration for softened peak interference power constraint. Our proposed scheme is able to combine the strength of both overlay and underlay transmission mode and provide superior performance through the use of the softened peak interference power constraint. In a cognitive radio system with dynamic primary user activity, the EBAS scheme is able to provide a much higher capacity to cooperative nodes compared to an overlay or underlay only approach. EBAS is also able to provide differentiated service to secondary users with different priority and to sustain the service differentiation in the presence of primary user activity. Although the physical and MAC layer solutions are able to improve the performance for the CRNs in terms of QoS provisioning, a judicious routing strategy is also equally important in ensuring a good performance in terms of the network delay and throughput. Thus, the second part of our research focuses on the design and implementation of dynamic routing protocols for CRNs. In the first part of our work on dynamic routing protocols, we designed and implemented a dynamic virtual carrier sensing and interference aware routing protocol (DVCSIR).The basic idea is that an optimum path is not necessarily the shortest hop one, but is one that has the least cost according to several criteria like interference, path length, traffic condition etc. The proposed DVCSIR scheme selects the optimum path based on two criteria, shortest path and Quietness Index (QI). The QI is a novel metric which we have developed to quantify the potential interference experienced by a node in the WMN. DVCSIR then makes use of the QI to find the least interference route from source to destination. We also introduced the dynamic virtual carrier sensing mechanism, where the virtual carrier sensing is performed on a node by node basis, depending on the node density. This allows the virtual carrier sensing mechanism to be applied only in parts of the network where collision is the likeliest to happen. The simulation results indicated that DVCSIR has superior performance due to its dynamic virtual carrier sensing mechanism, and interference aware path routing. In the second part of our research on routing protocols, we extend on the idea of multi-metric routing by proposing and implementing the Opportunistic Service Differentiation Routing Protocol (OSDRP) for the dynamic CRNs. OSDRP discovers the minimum delay – maximum stability route in dynamic CRNs by considering the availability of spectrum opportunity in addition to switching delay and queuing delay across primary user networks. In addition, service differentiation can be achieved through a combination of transmit power control and opportunistic routing. Simulation results demonstrate that the OSDRP achieves much better performance in terms of lower delay and higher throughput compared to other existing routing protocols across various scenarios. In the final part of our research, we consider the issue of security provisioning in CRNs. The designs of network protocols, especially those for spectrum sharing and management have introduced significant security loopholes. At the same time, most existing cognitive routing protocols assumed that each node honestly participates in packet forwarding. However, this assumption is no longer legitimate for CRNs due to the lack of a centralized trusted authority. QoS provisioning may also suffers as selfish nodes might incline towards self-centred behaviours in order to maximize their interests. Examples of such acts include deliberately dropping others’ packets, refusing to engage in forwarding of others’ packets or mislabelling all packets from the current node as high priority. These selfish behaviours can lead to monopolization of the wireless channel by the selfish nodes, or in a more severe case, the failure of routing protocols. To address the QoS provisioning issue in CRNs with selfish nodes coexistence, we propose a cross-layer Altruistic Differentiated Service Protocol (ADSP) for the dynamic CRNs in the presence of selfish nodes. Simulation results demonstrate that the ADSP can achieve much better performance in terms of lower delay, higher throughput and better delivery ratio for the traffic originating from collaborative nodes compared to other cognitive routing protocols in the presence of selfish nodes. We conclude our work and also provide a direction for future research at the end of the thesis.
DRNTU::Engineering::Electrical and electronic engineering::Wireless communication systems