Identification of spectrum holes using ANN model in TV bands with AWGN

Here we propose an artificial neural network (ANN) model for spectrum sensing in TV band specifically for detecting the presence of audio signals. The ANN model is trained with parameters which are a combination of cyclostationary and SNR based features like channel capacity, bandwidth efficiency, autocorrelation. The ANN model is trained with a new decision making factor termed as utilization factor (U) based on the above combination of attributes which lead to a method for detection of spectrum holes. The bandwidth efficiency (η) is also considered as a decision making factor to identify spectrum holes. This unique combination of hypotheses tries to remove the disadvantages of conventional energy detection and cyclostationary feature detection technique commonly used for CR applications.