Missile target type identification with Bayesian network

Features of missile target type identification were gradually sampled by the satellite sensors in the space-based early warning system, and the acquiring sequence exhibited random attribute but incomplete independency. Thereby, inference model based on Bayesian network was built, which could deal with the uncertain inference problem under feature random and feature incomplete independency condition. Moreover, the model was easily integrated into expert knowledge. Aiming at the problems of sampled data were not completely reliable due to many uncertain factors, such as the limitation of sensor ability, the impact of environment noise, the Bayesian network interfence model with uncertain evidence was established based on entropy gain. Simulation experiment showed that the Bayesian network model with belief inference impoved the identification precision by 7.35%.