Aerial Target Intention Recognition Based on Deep Belief Network

In view of the current complex battlefield data, this paper studies the objects and elements of tactical intention recognition in order to obtain the characteristics of typical aerial target. On this basis, a method for recognizing tactical intentions of aerial targets is constructed. Considering the difficulty of identifying with traditional expert knowledge and experience which is prone to different understandings, a deep belief network is used to recognize tactical intentions based on data learning and probabilistic reasoning methods. By introducing time factor, the proposed dynamic Bayesian network gives the process of event reasoning more continuity to get in line with reality. A genetic algorithm is used to obtain the Bayesian network structure based on a simulated data set, and then the Bayesian network is used for intention reasoning. Finally, a simulation test environment is built and a digital simulation is conducted to verify the effectiveness of the proposed method.