Premium
A new feature extraction and attacking node classification framework using PDCFE‐HPSMM model in cognitive radio networks
Author(s) -
Ramasamy Manjith
Publication year - 2016
Publication title -
international journal of communication systems
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.344
H-Index - 49
eISSN - 1099-1131
pISSN - 1074-5351
DOI - 10.1002/dac.3153
Subject(s) - computer science , cognitive radio , computer network , wireless , wireless network , hidden markov model , node (physics) , bandwidth (computing) , distributed computing , artificial intelligence , telecommunications , structural engineering , engineering
Summary The increasing demand of wireless communication introduces a challenge to an effective spectrum utilization. In wireless communication systems, the Cognitive Radio (CR) has emerged as a new key technology to address this issue, which allows an opportunistic access to the spectrum. The CRs form a CRN by extending the radio link features to network layer functions. The existing CR based wireless communication techniques has some drawbacks such as, inability to differentiate interference or noise from primary signals, it does not distinguish the different types of noise, low performance, not effective in spread spectrum detection, long observation time and high computational cost. To overwhelm these disadvantages, a novel method, namely, partial distribution based cyclostationary feature extraction–hidden probability state Markov model (PDCFE‐HPSMM) is proposed in this work. The main intention of this technique is to extract the node features, classify the attacker nodes and predict the licensed and unlicensed bandwidth. At first, the CRN is formed with some set of CR nodes, and the communication links between the nodes are estimated. Then, the node features are extracted by employing the PDCFE technique. After that, the attacking nodes are classified based on those features with the help of a HPSMM classification technique. If the attacker is present in the network, the link between the attacking nodes to other nodes is disconnected; otherwise, the communication link is updated. In this work, the Rayleigh channels are used to predict the licensed and unlicensed bandwidth. Again, the HPSMM model is employed to optimally select the channel, and the scheduling technique is implemented to check the status of the channel for message transmission. If the channel is not busy, the message signals are multiplexed and transmitted to the receiver via the selected channel. The novel concept of this paper is, the proposed PDCFE technique extracts the features of the nodes in CRN; based on these features, the attacker nodes in the network are classified with the help of HPSMM technique. Moreover, the HPSMM is used to select the optimal channel for message transmission. The experimental results evaluate the performance of the proposed system in terms of bit error rate (BER), error rate, relay, false detection rate (FDR) and cumulative distribution function (CDF). Copyright © 2016 John Wiley & Sons, Ltd.