Abstract
Existing mission abort policies for systems operating in random environments typically use the number of experienced shocks as a key decision parameter and assume that the shock detection mechanism is perfect. In practice, the shock monitoring system is failure prone, leading to wrong detections including both false negative (real shocks are not detected) and false positive (non-existent shocks are flagged) detections. This work advances the state of the art by modeling mission success probability (MSP) and system survival probability (SSP) of a system subject to imperfect shock detections and mission aborting. A dual-parameter abort policy (AP) is considered, which triggers the mission abort and starts a rescue procedure to survive the system when a predefined number of shock detections take place before a predetermined time. The AP optimization problem is solved to minimize the expected mission losses, balancing MSP and SSP. A case study of an unmanned aerial vehicle performing a surveillance mission in shock environments is provided to demonstrate the proposed model. The influences of shock detection errors and several other model parameters related to shock resistance, shock rate and costs on mission metrics and optimal APs are also investigated, leading to concrete managerial recommendations for controlling failure risks.