CYCLOSTATIONARY DETECTION USING FRESH FILTER FOR SPECTRUM SENSING IN COGNITIVE RADIO

Abstract

Radio spectrum has been instrumental in developing for itself as the most valuable resource in the field of wireless communication. Along with the networking based on cognitive features, the cognitive radios are being used as a measure for transmuting the fixed spectrum allotment in to spectrum allocation that is dynamic in nature. Cognitive radios are known as instruments which have infused intellect that makes them able to judge conditions of the environment. They can amend the features and parameters as per the need to obtain the optimal functioning at the nodes whether individual or when deployed at network level. In the electromagnetic spectrum they have been started to be considered as a new approach to improve upon the effective utilisation of a treasured natural asset. Cognitive Radio (CR) can by itself find free channels in a wireless spectrum. It is able to amend transmitting parameters which enables more communications to run simultaneously. It is termed as a radio which is smart and adapts itself to improve the behaviour of radio functioning. CR can be visualised as a goal, which should be attained by none other than the erstwhile software-defined radio: wireless transceiver that can be reconfigured and which by itself changes its communication parameters to cater for demand of the network and users. Presently, cognitive radio is of much interest to engineers because of importantly increasing the overall usage of spectrum. This is still a very saucy and interesting study topic, hence many technical research questions are required to be answered. To maintain itself as non-invading, a CR must follow certain rules strictly in the quality to sense the spectrum for primary users. In cognitive radio network a mighty role is played by spectrum sensing. The field has witnessed a number of algorithms which have been put forward with an aim to surge the phenomenon of spectrum sensing. Energy detection, matched filter detection and the cyclostationary detection pair up to become amongst the important algorithms for the same. IEEE 802.22, being the first standard in the domain that comprises of devices viz. CR devices, ensures very stern necessity in respect of probability of detection and the false alarm rate on devices performing spectrum sensing for SNR till a level of -20 dB. This needs improvisation of spectrum sensing methods which are robust in nature. ii Energy detection can be termed as the most simple and almost close to the most desirable technique which is over a great extent used in spectrum sensing. But somehow, the results are affected at lower values of SNR by uncertainty in noise variance. Similarly, in case where the primary user signal is known to the secondary user, matched filter detection technique provides the optimal output. In order to be able to detect signals which are characterised by low Signal to Noise Ratio, the spectral correlation can be exploited using cyclostationary based detection. The concept of frequency shift (FRESH) filter is incorporated, which consists of branches of frequency shifts applied to the received signal further passed through a linear filter. This causes suppression of the AWGN and also to the interference present in the received signal. The result gives an improved SNR value in the signal, which is then subjected to spectrum sensing using Energy Detection, Matched Filter Detection and Cyclostationary Detection techniques. The results show that the use of FRESH filter, gives significant improvement in the performance of all the mentioned spectrum sending techniques. This report covers the basics of cognitive radio, the concept of Spectrum Sensing, Energy Detection vis-à-vis Cyclostationary Detection and Matched Filter Detection before and after use of FRESH filter, their advantages and drawbacks and finally the future scope of work