This chapter presents a novel methodology for detecting logic bombs hidden in unmanned aerial vehicle autopilot code without source code analysis by executing mission runs in a software-in-the-loop simulator and defining safe unmanned aerial vehicle operating areas. The methodology uses preplanned flight paths as a baseline, greatly reducing the input space that must be searched to have confidence that an unmanned aerial vehicle will not encounter a triggering condition during its mission. While the focus is on detecting logic bombs in ArduPilot autopilot software, the methodology is general enough to be applicable to other unmanned aerial vehicle systems.