AbstractWhile increasingly popular, small unmanned aerial vehicles, aka drones, are often flown illegally over outdoor public gatherings or public facilities like prisons, threatening the safety of those nearby. There is a clear need to address interloping drones in public spaces from a sociotechnical perspective, including understanding the tradespace of variables. Through surveys, interviews, technology and infrastructure design, and experimentation, we developed a tradespace model of those variables that managers and designers of high-risk settings like public spaces and prisons need to consider in their development or renovation. These include cost considerations, both capital and infrastructure, as well as technology design elements of range and false alarm rates potentially exacerbated by convolutional neural networks (aka, deep learning). Results also highlight that environmental design elements can provide a possible low-tech solution in the design of obstructions that either eliminate or complicate a drone pilot’s line of sight. This effort demonstrates that managers and designers of high-risk settings like public spaces and prisons will have to balance sometimes competing objectives to obtain the best possible outcomes for public safety.
Learn to Path: Using neural networks to predict Dubins path characteristics for aerial vehicles in wind
