Ilmatorjunnan ilmasta-maahan -aseeseen aiheuttaman uhkan arviointi perustuen ilmatorjunnan ryhmityksen todennäköisyysmalliin

摘要

As a part of planning an air mission, the trajectory of an air-to-ground (A/G) weapon must be determined. In this thesis, the novel trajectory evaluation framework with which the best trajectory can be identified from a set of possible trajectories under uncertainty regarding the locations of surface-to-air (S/A) threats is presented. The best trajectory is the trajectory which has the highest survivability, i.e., the probability for the A/G weapon to traverse the trajectory without being intercepted.The trajectory evaluation framework relies on two new models introduced in this thesis which together provide the survivability of a given trajectory. The spatial prediction model is used to build a probability map for the location of an S/A threat based on Bayesian reasoning with geographical data and knowledge about common tactical principles utilised in forming an air defence. The Markov survivability model describes the process of intercepting an A/G weapon with the air defence consisting of radar sensors and S/A weapons with an inhomogeneous continuous-time Markov chain. Using the probability maps produced by the spatial prediction model, the Markov survivability model produces the survivability of the trajectory, such that uncertainties regarding the locations of the S/A threats are taken into account.The Markov survivability model presented in this thesis is compared with existing reference survivability models through numerical experiments by replacing it in the framework with each of the reference models. In the experiments, the survivabilities of different trajectories obtained with each model are evaluated and compared. The sensitivity of the models to uncertainty regarding the locations of S/A threats is studied by varying sizes of areas in which it is believed that the threats are located. The experiments imply that the novel framework gives intuitive results. In addition, the Markov survivability model is less affected by imprecise information regarding the locations of the S/A threats than the reference models.