UAVs have incredible potential for a variety of practical applications.  However, multiple hurdles lie in the way of everyday use cases for UAVs.  The Machine Games project will explore solutions to these problems in a real operating environment using our recently-completed ARENA (Autonomous Robotics Exploration Netted Arena).  Our goal is to bring drones into everyday life but there is a bevy of challenges which must first be overcome.  We identify three major problems facing mass-market adoption of drones:

1)  There is a need for a method via which a lone individual can control a large network of autonomous vehicles.  Current solutions to this problem either place too many burdens on the operator or allocate most of his control to a computer—either the computer or the operator are being used to his/her fullest, never both.  We hope to use the ARENA to study ways in which an operator and computer can cooperatively control autonomous systems.  One of our goals is to ensure that this mechanism makes sense at an intuitive level for an untrained user—a necessary step for drones to achieve mass-market success.

2)  Many studies, such as the Mesicopter project, have noted that existing techniques for both estimation and control are insufficient for the needs of everyday drone use.  Controllers may perform well in highly constrained environments but lab performance does not translate well to more realistic operating conditions.  Existing manners of performing position and attitude estimation are also insufficient for mass-market use, either due to inaccuracies in sensing or unreasonably high costs.  Machine Games will investigate and tackle both the control and estimation problems which stand in the way of everyday use.

3)  There exists a need for a defense system which can take down a rogue drone with limited risk to bystanders, infrastructure, or the system itself.  This lack was highlighted in January 2015 when a pilot lost control of a drone which crashed into the lawn of the White House.  At the time, no systems for downing rogue drones were in place and there is still a lack of a system which can safely remove UAV threats near dangerous or high-risk areas, such as airports.  Tactics for counter-UAV include spoofing, cannon-launched nets, and trained attack birds, though these solutions each have major flaws.  We hope to study methods for counter-UAV operations which can remove a rogue aircraft with limited risk to bystanders or infrastructure.

     Through the course of the life of the ARENA, we will investigate a variety of open questions in security and autonomy which stand in the way of mass-market adoption of UAVs.  Many of the challenges we investigate will have impacts outside of the field of drones; for example, questions of position determination are relevant to autonomous cars.  The solutions we develop to these three problems will help guide legislation to promote drone use without compromising safety, will drive entrepreneurs to invest into otherwise dangerous propositions, and will encourage the use of UAVs as a solution to commonplace problems.

Watch us on KXAN