Search-and-rescue operations, package delivery, and underwater exploration could all be performed soon by intelligent machines. The Autonomy Incubator group at NASA Langley is taking strides to, as group leader Danette Allen puts it, "imbue machines with the kind of intelligence that we expect from human beings."

Given a simple voice command to take off, this small quadcopter can rise vertically from a home base, zero in on a barrier marked with a target, fly up to the barrier, land, then take off again and return to home base. (NASA/David C. Bowman)
Allen and her group believe the research they're doing could lay the groundwork for a future in which NASA can safely and reliably fulfill that need. Some of the areas that would benefit from autonomous systems are deep space exploration, as well as unmanned and personal air vehicles. In order to make possible a future in which unmanned and personal air vehicles can safely fly the nation's skyways, engineers must design systems that will allow the air vehicles themselves to handle some of the complex decision making of flight.

Allen defines true autonomy as the complete delegation of authority and decision-making, "not unlike telling your dog to fetch the paper. You don't tell them how to do it. You don't set any settings. It's up to your dog to decide what trajectory to follow, how to bring it back." An autonomous system will learn from its experiences in the same way that a human does and, with that experience, improve its performance, expand its knowledge base, and make decisions in the face of uncertainty.

Source 

Topics:
Aeronautics Aviation Data Acquisition Data Acquisition Machinery Sensors

More From SAE Media Group