TY - JOUR
AU - Бугайов, М. В.
PY - 2022/12/30
Y2 - 2023/03/31
TI - Method of Optimal Test Statistic Search for Spectrum Sensing
JF - Visnyk NTUU KPI Seriia - Radiotekhnika Radioaparatobuduvannia
JA - RADAP
VL -
IS - 90
SE -
DO - 10.20535/RADAP.2022.90.13-20
UR - https://radap.kpi.ua/radiotechnique/article/view/1786
SP - 13-20
AB - <p>The constant growth in the number of electronic devices leads to increasing of spectrum occupancy. In addition, new data transmission technologies are constantly used and time-frequency structure of signals become more complicate. These factors lead to a significant complication of electronic environment, which leads to new approaches to fast spectrum sensing.<br /><br />In most publications, to make a decision about the presence or absence of signals in a given frequency band, some transform is taken from the signal, followed by the calculation of test statistics. However, it is not indicated for what reasons this type of test statistics was chosen and whether it is optimal for a given transform and type of signal spectrum shape.<br /><br />Problem of choosing the optimal type of test statistics is especially actual when working in conditions of unknown and variable noise power, as well as with a wide dynamic range of signals. Test statistics should be sensitive to spectrum outliers. The essence of the proposed method is forming a set of test statistics and calculating the value of coefficient of efficiency as the sum of detection probabilities for different spectrum shapes using these statistics for different signal-to-noise ratios and spectrum occupancy. The maximum value of the coefficient of efficiency will correspond to the optimal type of test statistics.<br /><br />As a result of the research, it was found that to separate frequency samples into signal and noise, it is advisable to use coefficient of variation. Prospects of further research in this direction should be focused on development of methods for dynamic transition between types of test statistics in process of radio monitoring, depending on changes in the electronic environment.</p>
ER -