Decision making method of weapon station in fully connected residual network

DOI:	10.11809/bqzbgcxb2023.12.031
Keywords:	weapon station decision making; nonlinear factors; residual network; batch normalization; discard layer
Abstract:	A weapon station is a new weapon system that can be operated remotely, equipped with multiple weapon combinations, and when the target is identified, the attack is carried out with the appropriate weapon. At present, most of the decisions on the use of weapons are judged by commanders based on human experience, they lack of autonomy, and cannot process various complex battlefield situation information. To solve the above problems, an autonomous decision making method of fully connected residual network weapon stations based on deep learning is proposed. Taking the main factors affecting the decision of the weapon station as input, based on the fully connected neural network, the batch standardized optimization algorithm, the discarded layer and the residual module (ResNet) are added to construct the model, the model output and sample labels are compared, the influence of the number of training samples on the accuracy of the model is analyzed, and the accuracy and loss changes of the models under the same test set under three different training data volumes are compared. Model verification and simulation show that the proposed method can make autonomous decision making on weapon stations.
Issue:	Vol. 44 No. 12 (2023)
Published:	2023-12-28
PDF HTML