monash_118403.pdf (1.46 MB)
Provisioning quality-of-services in cognitive radio networks
thesisposted on 2017-02-17, 01:58 authored by Hasan, Md Rashidul
To meet the growing demand for bandwidth and to improve the utilisation of frequency spectrum, cognitive radio (CR) evolves as the key enabling technology, which allows CR users (or unlicensed users) to access licensed spectrum opportunistically. Accessing the spectrum in an opportunistic way introduces many research challenges, including exploiting the behaviour of spectrum holes, sharing spectrum holes by multiple access techniques, minimising interference and enforcing a semi-formal agreement between CR users and licensed users. Additionally, supporting delay-sensitive real-time services through CR networks, so that unlicensed users are allowed to experience full-scale networking services by opportunistically sharing spectrum from a number of existing networks without impacting on the licensed users, remains a challenging task. To address these challenges, CR users require an efficient and effective medium access control (MAC) framework, which will not only ensure quality-of-services (QoS) in the presence of resource uncertainty but also guarantee an agreed level of interference with licensed signals. This thesis aims at provisioning QoS in CR networks by proposing a new MAC framework that introduces a novel interference minimisation technique by exploiting the usage characteristics of licensed signals. To accommodate the surging demand for bandwidth, we have investigated whether the general purpose CR ad hoc networks are capable of fulfilling QoS through opportunistic spectrum sharing. By empowering CR users with multiple channel aggregation techniques, we have drawn a useful limit between the hardware capability and the channel availability for certain QoS requirements. To meet the requirements of an agreed interference level, we have proposed an optimal transmission holding time by CR users for each licensed channel by capturing the dynamic behaviour of licensed users. The proposed MAC accommodates multiple CR users in a multichannel scenario and ensures fair load balancing across the channels. The applicability of the proposed MAC is ascertained by thorough analysis of the network related QoS parameters, i.e., throughput and delay. Our proposed MAC framework eventually ensures non-intrusive co-existence of CR and licensed users, which is the pre-condition for wide acceptability of CR technology.