Özet:

Short radio waves propagate in the ionosphere - randomly-inhomogeneous medium of the upper atmosphere. Gain control system is one of the important parts of a short-wave radio receiver. Frequently the received signal has a wide dynamic range. A conventional approach for this problem consists of using an automatic gain control. In that case the receiver's gain is automatically adjusted to maintain a constant level using averaging filters with different time constants. It is enough in most communication tasks. Solving tasks of radio waves propagation or ionospheric parameters investigation, it is necessary to predict mean level of radio signal absolute value of the received signal. In this case an adaptive gain control system can be useful. We proposed the method of developing self-learning gain control system. Self-learning neural network is used to estimate mean value of the signal amplitude. Input vector of the neural network consists of previous samples of the signal, the network output is estimated gain level for next several samples. Since short-wave radio receiver is a self-contained device, the self-learning gain control system based on the neural network was realized on programming logic device.

Anahtar Kelimeler: