Predictive missile guidance for agile maneuvering targets with stochastic hybrid dynamics

In this work we consider the problem of designing predictive guidance and control laws for missiles targeting agile aircraft. The target and missile dynamics are modeled as stochastic hybrid systems, which enables expression of complex maneuvers as sequences of simpler maneuver modes. Target's missile evasion strategy is modeled based on domain expertise and the interception problem is formulated as a Partially Observable Markov Decision Process, which is approximately solved online. The results show that the developed guidance method is superior to methods that do not utilize prediction of target's maneuvers or ignores the uncertainty in target's evasion tactics.