TY - GEN
T1 - New proposal for eliminating interferences in a radar system
AU - Campa, Carlos
AU - Acevedo, Antonio
AU - Acevedo, Elena
N1 - Funding Information:
The authors would like to thank the Instituto Politécnico Nacional (COFAA and SIP), and SNI for their economical support to develop this work.
PY - 2010
Y1 - 2010
N2 - 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.
AB - 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.
KW - Backpropagation
KW - FIR low-pass filter
KW - Noise elimination
KW - Radar system
UR - http://www.scopus.com/inward/record.url?scp=78649996513&partnerID=8YFLogxK
U2 - 10.1007/978-3-642-16773-7_38
DO - 10.1007/978-3-642-16773-7_38
M3 - Contribución a la conferencia
SN - 3642167721
SN - 9783642167720
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 440
EP - 453
BT - Advances in Soft Computing - 9th Mexican International Conference on Artificial Intelligence, MICAI 2010, Proceedings
T2 - 9th Mexican International Conference on Artificial Intelligence, MICAI 2010
Y2 - 8 November 2010 through 13 November 2010
ER -