From "dazzling flying machines" to "autonomous killer drones" ...

A little more than a decade ago, I was working as a scientist and principal investigator at the University of Sydney where I initiated the "Slow Flyer" project. My research interests were focused on the design of bio-inspired vision systems for Micro Aerial Vehicles (MAVs), and more particularly self-motion estimation from a single camera, wide-angle depth perception, and computational catadioptric cameras.

While digging through some files on my backup drive, I came accross an old presentation [see below figure] that illustrates my visionary thoughts on the sensing requirements for MAVs at the time ... Examples of target applications included space exploration and urban counter-terrorism.

In early 2016, I was pleasantly surprised to see that drones were used in many non-military and non-defense-related applications such as filming/aerial photography, search and rescue missions, geographic mapping, precision agriculture ... In particular, I truly got impressed by the work of Raffaelo D'Andrea, and I liked to dream about the idea that a swarm of small drones could have aesthetic value.

Well, since late 2017, I started to realise that all perception and navigation capabilities are now available thanks to the rapid advances in sensing technologies and decentralized artificial intelligence based on deep learning machines, powered by neural netwroks and trained with big data sets. Nevertheless, I never expected that the trajectory of those MAVs tech's developments could turn to the dark side so quickly ... Check the below video on autonomous killer drones to be used as weapons of mass destruction!