Вісник НТУУ «КПІ». Радіотехніка, радіоапаратобудування : збірник наукових праць, Вип. 94

Постійне посилання зібрання

https://ela.kpi.ua/handle/123456789/65632

Переглянути

Перегляд Вісник НТУУ «КПІ». Радіотехніка, радіоапаратобудування : збірник наукових праць, Вип. 94 за Ключові слова "621.3:519.2"

Зараз показуємо 1 - 1 з 1
  • Ескіз
    ДокументВідкритий доступ
    Energy detector of stochastic signals in noise uncertainty
    (КПІ ім. Ігоря Сікорського, 2023) Buhaiov, M. V.
    Wide use of software-defined radio has led to a significant sophistication of electronic environment. This is mainly due to ability of generation signals of almost any shape. To detect signals with an unknown dynamic frequency-time structure, it is advisable to use advanced energy detector algorithms. The purpose of this article is to automate processes of stochastic signals detection and time parameters estimation under the conditions of unknown frequency-time structure of signals and noise power. The essence of proposed method is to detect and track temporal energy changes averaged over L samples of received signal in selected frequency channel. Threshold value for a given probability of false alarm is calculated using current estimates of signal power. This threshold is dynamic and is refined only in time intervals free from the signals. In those time windows where energy exceeds threshold, a decision is made about the presence of a signal. An algorithm for detecting stochastic signals is proposed. If a signal is present at the initial moment of time, proposed algorithm can detect only its end by a sharp decrease of signal energy. After that, new noise level is estimated and threshold value is refined. Detection curves of proposed algorithm are obtained. It is shown that when number of samples L is increased by an order, the gain in signal-to-noise ratio in signal detection is about 4 dB. The maximum value of correct detection probability of a pulse signal is achieved with the same pulse duration and the length of the integration interval. Compared to method of signal smoothing with moving average window, proposed method has less computational complexity, simplifies the search for signal time boundaries, and gives smaller errors in signal duration estimates. Recommendations for the implementation of developed algorithm are formulated.