This paper addresses the dynamic mission reliability evaluation of unmanned aerial vehicle (UAV) swarms. To handle the inherent complexity of swarm systems and the limitations of existing static evaluation methods, a new approach based on an improved dynamic Bayesian network (DBN) is proposed. A hierarchical hidden state space is constructed on the basis of the DBN, and a nonlinear state transition model is employed to capture the coupling effects among performance indicators as well as the associated degradation patterns. Stage-dependent observation models are then integrated with measurement data, and particle filtering is used to perform online state estimation. Finally, the mission reliability of the UAV swarm is derived from the estimation results. Simulation case studies demonstrate the effectiveness and feasibility of the proposed method.
Liu et al. (Fri,) studied this question.