On Wednesday the 26th of Aug. 2020, 09:00 AM, the Ph.D. student “Muntaser Salim Faleh“defended his thesis entitled:

” Implementation of Improved Sensing and Sharing schemes for Cognitive Radio System “

The discussion committee included:

Prof. Dr. Nasr N. Khamis / College of Information Engineering / Chairman.

Prof. Dr. Raad H. Zahir / Al-Mustansiriya University / College of Engineering / Member.

Prof. Dr.Hadi T. Zeboon /University of Technology / Electrical Engineering Department / Member.

Prof. Dr. Ali S. Abdel-Hadi / College of Information Engineering / Member.

ِAssist.Prof.Dr. Wael A. Hadi / University of Technology / Department of Communications Engineering/member.

Prof.Dr. Hikmat N. Abdullah / College of Information Engineering / Member and supervisor.

The discussion was also attended by the head of the department, a number of faculty members, Ph.D. students in the department, the researcher fulfilled the requirements for obtaining a Ph.D. degree with a degree of (excellent).


Abstract :

Cognitive Radio (CR) is an emerging and promising technology to solve the sacristy problem in the spectrum and it is the future of 5G communication systems. CR is an intelligent system that aims to manage the available spectrum in an efficient manner. The most sensitive stage in CR life cycle is Spectrum Sensing (SS). Energy Detection (ED) based SS, is the most preferred sensing scheme due to its simplicity. It must be enhanced to overcome some challenges. These challenges include noise uncertainty at confusion region, long sensing period, and weak detection performance at low SNR values especially for Multi-Band Spectrum Sensing (MBSS) and Wide-Band Spectrum Sensing (WBSS) scenarios. In addition, to satisfy the imposed SS requirements of IEEE.802.22 standard with affective sensing time and throughput.

In this thesis, six algorithms are proposed to enhance the performance of ED-based SS in the presence of the challenges and requirements mentioned above. The proposed algorithms are evaluated in both AWGN and Rayleigh fading channels scenarios through MATLAB simulations. Also, they are validated using real-time spectrum data and realized in the form of embedded systems using SDR devices. The proposed algorithms are divided into three parts:

 The first part is concerned with proposing efficient decision mechanisms in the confusion region when the target signal falls in the intersected region between Primary User (PU) and noise limits. In this context, two Double Threshed Energy Detection (DTED) algorithms are proposed. The first algorithm, called Modified Double Threshold Energy Detection (MDTED) is based on dividing the confusion region into sub-regions and taking a set of decisions. The second algorithm, called Knowledge-Based Decision Procedure (KBDP), is designed to use present and previous testing as test statistics. The simulation results of this part show that at SNR= -12dB, the Detection Probability (PD) in the AWGN channel is increased by ratios 19% and 20% using the MDTED algorithm  KBDP as compared to the other DTED algorithms, respectively.

The second part is targeted to develop hybrid SS schemes to meet MBSS requirements. In this context, the useful features of Discrete Cosine Transform (DCT) and Discrete Wavelet Transform (DWT) are utilized for recognizing PU from noise only signals. Three algorithms are proposed in this part. The first two algorithms are for Uncooperative Spectrum Sensing (USS) named AR-DCT and AR-DWT algorithms, while the third one is for multi-stages Cooperative Spectrum Sensing (CSS) called DWT-DCT-CSS algorithm. The simulation results show that, at SNR=-12dB and in the AWGN channel, the AR-DCT algorithm enhances PD by ratio 115% compared to CED and Coefficient Variation (CV) approaches. Regarding the AR-DWT algorithm, PD has increased the ratio of 85% with respect to CV algorithm. Finally, in the DWT-DCT-CSS algorithm, PD is significantly improved when the number of Secondary Users (SU) equals 10 compared with traditional hard fusion rules.

The third part addresses SS in WBSS based on non-reconstructive Compressed Spectrum Sensing (Comp SS). In this context, a novel SS algorithm is proposed using Maximal Overlap Discrete Wavelet Transform (MODWT), it is called MODWT SS algorithm. The simulation results show that, in AWGN channel scenario, at SNR=-19 dB MODWT SS algorithm has increased PD by the ratio of 98% compared to the proposed method.

Finally, the proposed algorithms are realized using an embedded system. The realized system includes using HackRF One SDR with GNU Radio as PU and RTL SDR with specialist GUI MATLAB profile as SU. The system realization is formulated in both USS and CSS using real-time spectrum data. The implementation results show that the proposed algorithms achieved good PU detection at low SNR values with closed performance to the simulated results.