New proposal for eliminating interferences in a radar system

In this work we present a new proposal to initialize the weights in a Backpropagation Neuronal Network (NN) using the coefficients from a FIR Low-Pass Filter to introduce a null in the radiation pattern in a seven-element array of antennas to eliminate interferences in a radar system. A radar system needs to eliminate the directional noise in order to obtain a cleaner signal. The method used to eliminate this kind of noise (jitter) has to be adaptive because the objective is in constant movement, therefore, the adaptation time must be as fast as possible. Our work is based on the window method to reduce the secondary lobes in fixed arrays of antennas. We modify the radiation pattern by introducing a null at 45.5° which corresponds to the secondary lobe where the interference is presented. This is achieved when we create windows from several FIR Low-Pass Filters. The coefficients of these filters are used to initialize the weight vectors of a Backpropagation Neural Network which performs the adaptive process to obtain the final parameters to achieve the noise elimination. For testing our proposal we calculate the Mean Square Error (MSE), the Signal Noise Relation (SNR) and we graphed the Radiation Pattern. In addition we calculated the Cross Correlation Index in each iteration, between the desired signal and our results. With this method we reduced the number of iterations required by the process.