{"id":90872,"date":"2017-05-30T06:00:00","date_gmt":"2017-05-30T11:00:00","guid":{"rendered":"https:\/\/www.premiumbeat.com\/blog\/?p=90872"},"modified":"2017-05-30T10:07:22","modified_gmt":"2017-05-30T15:07:22","slug":"mit-autonomous-aerial-cinematography-drones","status":"publish","type":"post","link":"https:\/\/www.premiumbeat.com\/blog\/mit-autonomous-aerial-cinematography-drones\/","title":{"rendered":"Future Tech: Peek Into The Future of Aerial Cinematography"},"content":{"rendered":"

MIT Researchers are developing a new software-based system for aerial\u00a0drone cinematography.<\/h2>\n

Cover image via Shutterstock<\/a>.<\/em><\/p>\n

The days of dual drone pilots and camera operators<\/strong> may be numbered. According to MIT News<\/a>, scientists at the Massachusetts Institute of Technology<\/strong> are currently developing a new software that allows drones to frame (and maintain) shots while performing aerial maneuvers<\/strong>\u00a0\u2014 while also avoiding obstacles).<\/p>\n

Autonomous Filming<\/h3>\n

\"FutureTech:<\/a><\/p>\n

\u201cThere are other efforts to do autonomous filming with one drone. They can follow someone, but if the subject turns, say 180 degrees, the drone will end up showing the back of the subject. With our solution, if the subject turns 180 degrees, our drones are able to circle around and keep focus on the face. We are able to specify richer higher-level constraints for the drones.” \u2014Daniela Rus, MIT<\/p><\/blockquote>\n

While self-flying drone technology already exists on the market<\/a>, MIT’s new software<\/strong> may be the beginnings of a far more autonomous and intelligent version<\/strong>\u00a0that can recognize more precise framing concepts and opportunities<\/strong>. As you can see in the examples above, the technology can recognize and program variables\u00a0such as shot size, viewing angle, and screen position<\/strong>.<\/p>\n

Viewpoint and Obstacle Avoidance<\/h3>\n

\"FutureTech:<\/a><\/p>\n

“The key to the system is that it continuously estimates the velocities of all of the moving objects in the drone\u2019s environment and projects their locations a second or two into the future. This buys it a little time to compute optimal flight trajectories and also ensures that it can get recover smoothly if the drone needs to take evasive action to avoid collision.” \u2014Javier Alonso-Mora, MIT<\/p><\/blockquote>\n

The new software<\/strong> also represents a major advancement in AI-recognized elements such as velocity and location<\/strong>. The drones can\u00a0“update its position projections about 50 times a second<\/strong>,” which allows for greater calculations and predictions<\/strong>\u00a0for where it needs to be to avoid collisions while maintaining viewpoints<\/strong>.<\/p>\n

An Open Future<\/h3>\n