Spectrum Sensing for Cognitive Radio Using Blind Source Separation and Hidden Markov Model Amrit Mukherjee 1 , Satyabrata Maiti 2 Research Scholar 1 , M.Tech 2 School of Electronics Engineering, KIIT University Bhubaneswar, India amrit1460@gmail.com, satyabratamaiti@yahoo.com Dr. Amlan Datta Professor and Associate Dean School of Electronics Engineering, KIIT University Bhubaneswar, India Abstract— Most of the radio frequency spectrum is not being utilized efficiently. The utilization can be improved by including unlicensed users to exploit the radio frequency spectrum by not creating any interference to the primary users. For Cognitive Radio, the main issue is to sense and then identify all spectrum holes present in the environment. In this paper, we are proposing the Blind Source Separation (BSS) sensing which is applied through the Hidden Markov Model (HMM). It does not need any kind of synchronizing signals from the Primary user as well as with the secondary transmitter in a working condition. Simulation results by the proposed method for BSS by the activity of Primary User (PU) have been presented. Index Terms—Cognitive Radio (CR), Blind Spectrum Sensing (BSS), Primary User Activity Prediction, Hidden Markov Model (HMM), Channel state prediction I.INTRODUCTION Wide applications of radio signals have resulted in obtaining a maximum usage of radio spectrum. The new factors coming into picture elaborate the need of spectrum management. Survey has shown about the actual utilization of spectrum, where it is clearly mentioned about the unused spectrum in the allocated range. These are reported by Federal Communications Commission (FCC)’s Spectrum Task Force. This introduced the need of Cognitive Radio (CR) with which we can improve the allocated spectrum efficiency. Here we can adequately use the unused spectrum in random and continuously changing environment. These involves in obtaining the different transmission attributes viz. power, latency, bandwidth of the signal, symbol rate of the different combination of the required potential users which depends upon the nature and behavior of the primary user. Spectrum sensing for cognitive radios is still an ongoing development and the techniques for the primary signal detection are limited in the present [1]. One of the most important and unique property of CR networks is the ability to shift and change between two different radio access technologies (ISM and Sensor networks) as idle and different frequency band slots arise [2]. This dynamic spectrum access which was proposed in [3] and is one of the most basic transmitter requirement to adapt to the criteria like varying quality of the channel, the available network congestion, channel to signal interference and service requirements of the channel. The secondary users of CR will also need to coexist with primary users (PU), as they are having the right to use spectrum and thus must have a surety not to be interfered by secondary users. Fig.1 shows the basic cognition cycle model, which focuses on spectrum processing. The cognitive capability of the system to allows it to interact with the nearby surrounding environment, and results in selecting proper communication parameters for that specific environment. Fig1. Cognitive Cycle The involvement of Blind source separation (BSS) in cognitive radio systems is introduced. BSS spectrum sensing methods are proposed [5]. The advantage of using BSS is that these can work without any synchronization for the primary signal from the transmitter in presence of the secondary transmitter working in an active mode. The multi-frequency spectrum sensing is implemented to distinguish the combined and mixed signals which are present in different frequency band [7]. Another attempt is made to separate the mixed and combined observed signals with or without the presence of PU which are based the auto and cross correlation between that different separated signals. The Energy Detection Techniques are mainly divided into two categories: “Transmitter Detection” and “Receiver Detection”. In the first category, the Primary User (PU) is assumed to be transmitting and in the latter PU is receiving. In this paper we have emphasized on transmitter detection. 2014 Fourth International Conference on Advanced Computing & Communication Technologies 978-1-4799-4910-6/14 $31.00 © 2014 IEEE DOI 10.1109/ACCT.2014.63 409