Study on Battle Damage Level Prediction Using Hybrid-learning Algorithm

Abstract

It is important to predict battle damage level timely and accurately for operation commander to adjust firing action intent, issue command, control situations, and make decisions correctly. Adaptive neural fuzzy inference system (ANFIS) architecture and the hybrid-learning algorithm by applying back-propagation and least mean squares procedure are studied. ANFIS model for battle damage level prediction is established based on the analysis of the main influence factors of battle damage level. The prediction of battle damage level being consistent with the factual damage level is achieved by training the proposed ANFIS model using damage test data. Simulations comparing analysis for battle damage level prediction results are conducted using the proposed method and BP neutral network respectively. Simulation results demonstrate that the proposed method can predict battle damage level correctly and the precision is higher than that of BP neutral network, and thus may provide an effective method for battle damage level prediction.