intelligence of electronic warfare offensive and deception systems using neural networks

Electronic warfare (ew) is one of the most important features of modern and contemporary battles. algorithms based on artificial intelligence (ai) can play a very effective role in various areas of ew, such as: processing radar signals to identify and classify transmitter types, detecting the type of jamming operation and its characteristics, as well as development, and having effective anti-interference algorithms. due to the non-linear nature of the patterns, it is not possible to identify and classify the data used in this article in a linear way. therefore, neural networks with linear structure and adaptive linear networks cannot be used. according to the array corresponding to the coded threats as the system input and the desired output of the network, which is an effective command in choosing the electronic countermeasure method, so we seek to identify the patterns of threat signals and countermeasure techniques in the classification stages, identification and coding to find a suitable method for classifying the input patterns and assigning the relationship between the input and output patterns. accordingly, neural networks multilayer perceptron (mlp), radial basis and that are competitive with lvq training rule to implement algorithms, make the arrays proposed in this paper desirable and a good option for problem solving and smartening of a particular electronic and telecommunication systems. based on the values obtained for the sensitivity, accuracy and especially the output responses for noise jamming and deception jamming techniques during the simulations carried out in this article, by changing the learning rate, applying the number of neurons with different iteration steps and different accuracies, the mlp neural network has a better condition than the other two networks. the mentioned neural network using traingda training function can perform deception jamming with 99.8% accuracy, 94.5% sensitivity, and 89.7% specificity and noise jamming with 99% accuracy, 91.86% sensitivity and 91% specificity to it has been achieved and therefore it is considered a suitable option for making ew offense and deception systems intelligent.

intelligence of electronic warfare offensive and deception systems using neural networks

electronic warfare (ew) is one of the most important features of modern and contemporary battles. algorithms based on artificial intelligence (ai) can play a very effective role in various areas of ew, such as: processing radar signals to identify and classify transmitter types, detecting the type of jamming operation and its characteristics, as well as development, and having effective anti-interference algorithms. due to the non-linear nature of the patterns, it is not possible to identify and classify the data used in this article in a linear way. therefore, neural networks with linear structure and adaptive linear networks cannot be used. according to the array corresponding to the coded threats as the system input and the desired output of the network, which is an effective command in choosing the electronic countermeasure method, so we seek to identify the patterns of threat signals and countermeasure techniques in the classification stages, identification and coding to find a suitable method for classifying the input patterns and assigning the relationship between the input and output patterns. accordingly, neural networks multilayer perceptron (mlp), radial basis and that are competitive with lvq training rule to implement algorithms, make the arrays proposed in this paper desirable and a good option for problem solving and smartening of a particular electronic and telecommunication systems. based on the values obtained for the sensitivity, accuracy and especially the output responses for noise jamming and deception jamming techniques during the simulations carried out in this article, by changing the learning rate, applying the number of neurons with different iteration steps and different accuracies, the mlp neural network has a better condition than the other two networks. the mentioned neural network using traingda training function can perform deception jamming with 99.8% accuracy, 94.5% sensitivity, and 89.7% specificity and noise jamming with 99% accuracy, 91.86% sensitivity and 91% specificity to it has been achieved and therefore it is considered a suitable option for making ew offense and deception systems intelligent.