Spectrum measurement, sensing, analysis and simulation in the context of cognitive radio

The radio frequency (RF) spectrum is a scarce natural resource, currently regulated locally by national agencies. Spectrum has been assigned to different services and it is very difficult for emerging wireless technologies to gain access due to rigid spectmm policy and heavy opportunity cost. Current spectrum management by licensing causes artificial spectrum scarcity. Spectrum monitoring shows that many frequencies and times are unused. Dynamic spectrum access (DSA) is a potential solution to low spectrum efficiency. In DSA, an unlicensed user opportunistically uses vacant licensed spectrum with the help of cognitive radio. Cognitive radio is a key enabling technology for DSA. In a cognitive radio system, an unlicensed Secondary User (SU) identifies vacant licensed spectrum allocated to a Primary User (PU) and uses it without harmful interference to the PU. Cognitive radio increases spectrum usage efficiency while protecting legacy-licensed systems. The purpose of this thesis is to bring together a group of CR concepts and explore how we can make the transition from conventional radio to cognitive radio. Specific goals of the thesis are firstly the measurement of the radio spectrum to understand the current spectrum usage in the Humber region, UK in the context of cognitive radio. Secondly, to characterise the performance of cyclostationary feature detectors through theoretical analysis, hardware implementation, and real-time performance measurements. Thirdly, to mitigate the effect of degradation due to multipath fading and shadowing, the use of -wideband cooperative sensing techniques using adaptive sensing technique and multi-bit soft decision is proposed, which it is believed will introduce more spectral opportunities over wider frequency ranges and achieve higher opportunistic aggregate throughput.Understanding spectrum usage is the first step toward the future deployment of cognitive radio systems. Several spectrum usage measurement campaigns have been performed, mainly in the USA and Europe. These studies show locality and time dependence. In the first part of this thesis a spectrum usage measurement campaign in the Humber region, is reported. Spectrum usage patterns are identified and noise is characterised. A significant amount of spectrum was shown to be underutilized and available for the secondary use. The second part addresses the question: how can you tell if a spectrum channel is being used? Two spectrum sensing techniques are evaluated: Energy Detection and Cyclostationary Feature Detection. The performance of these techniques is compared using the measurements performed in the second part of the thesis. Cyclostationary feature detection is shown to be more robust to noise. The final part of the thesis considers the identification of vacant channels by combining spectrum measurements from multiple locations, known as cooperative sensing. Wideband cooperative sensing is proposed using multi resolution spectrum sensing (MRSS) with a multi-bit decision technique. Next, a two-stage adaptive system with cooperative wideband sensing is proposed based on the combination of energy detection and cyclostationary feature detection. Simulations using the system above indicate that the two-stage adaptive sensing cooperative wideband outperforms single site detection in terms of detection success and mean detection time in the context of wideband cooperative sensing