Cyber risk loss distribution for various scale drone delivery systems

Thousands of commercial drone deliveries are occurring daily worldwide, mainly in the U.S. and China. Global corporate efforts such as the recent Amazon trials at College Station, TX, authorized by the FAA, are poised to make this activity ubiquitous. The potential various-scale applications will lead to substantial cyber risk exposure of deployed drone networks leading to cyberattacks such as hijacking and signal disruption, as evidenced by recent military applications. Thus in the commercial realm, these attacks may result in substantial financial losses from costly property damages. The limited availability of relevant historical losses presents a significant challenge in pricing these risks. Thus, we propose a structural model of an aggregate loss distribution for the cyber risk of drone delivery operations on a random spatial network. We contextualize the problem in the probabilistic graph-theoretical framework using bond percolation models and simulated spatial point processes. By characterizing the loss distribution, we allow for the further utility of pricing cyber risk for small, medium, and large-scale drone fleets.