Multiserver queueing for supervisory control of autonomous vehicles

Widespread adoption of unmanned aerial vehicles has led to intense interest in increasing the vehicle-to-operator ratio in unmanned systems. We propose a multiserver queueing structure for the supervisory control of multiple autonomous unmanned vehicles and demonstrate that multiserver queueing can increase the vehicle-to-operator ratio. Analytical results are derived for M/M/c queues with identically and non-identically distributed service times, and vehicle-to-operator ratio increases of up to 50% are predicted for multiserver queues compared to single server queues. A human fatigue model is applied to determine the performance of fixed shift and staggered shift structures in the presence of operator fatigue. Results are extended to log-normally distributed service times in discrete event simulation.