Abstract
Reliability is a pivotal requirement for drones or unmanned vehicles-based mission systems. Despite significant research efforts on their reliability modeling and analysis, the existing models are mainly suitable for single-drone systems or homogeneous multi-drone systems that perform multi-phase missions. The heterogeneous and collaborative requirements demand new reliability analysis methods for the drone-based multi-phase mission systems. This paper proposes two alternative analytical models to consider heterogeneity in robots executing the mission as well as collaboration and dependence between different phases. As demonstrated by a case study of a two-phase rescue mission system performed by drones with the exponential and Weibull time-to-failures, the proposed reliability analysis models are flexible in handling diverse types of drone time-to-failure distributions. Based on the case study, effects of several model parameters on the mission reliability are also examined. It is confirmed that as the scale parameter of the Weibull distribution or the failure rate under the exponential distribution increases, the mission reliability decreases. The impact of the shape parameter under the Weibull distribution depends on the mission time; the mission reliability increases and then decreases with the increasing shape parameter as time proceeds.