| Summary: | According to a survey by the Federal Communications Commission, a US government agency, the spectrum allocated to paying users is underutilised in time and space. In other words, there are still many opportunities for others to make use of these gaps to utilise the network. Spectrum sensing is about finding out if a spectrum is being used. By measuring certain properties of the received signal and comparing it to certain benchmark, it can be determined if any primary user (PU) is using the spectrum. If the PU is not using the spectrum, a secondary user (SU) can take the opportunity to use it. In this report, the brief concept of different spectrum sensing techniques is obtained from research papers. In particular, energy detection is chosen to be further explored. From the readings, the energy detector’s performance is affected by additive white Gaussian noise (AWGN), fading and shadowing channels. However, when cooperative sensing is used, their performance improves. Next, a simulation of the energy detector is done under an AWGN channel. It is done on the MATLAB software. The system model based on a research paper is used, and a code is written on it. The code compares the simulated and theoretical energy detectors, by drawing various graphs for both of them. They include probability of detection against probability of false alarm, probability of false alarm against threshold and probability of detection against threshold. Graphs are drawn for each of them, one with fixed signal-to-noise ratio but different time-bandwidth product, and another with fixed time-bandwidth product but different signal-to-noise ratio. Increasing the time-bandwidth product worsens the performance while increasing the signal-to-noise ratio improves the performance. It is observed that the probability of false alarm and detection decreases as the set threshold increases. This property can be used to design energy detectors for different uses. Though only energy detector was studied in depth, matched-filter detection and cyclostationarity feature detection are also potential research topics.
|
|---|