Perched landings afford aerial craft several distinct advantages, namely the ability to land in much smaller and more confined spaces, an increased capability for intelligence-gathering and a new degree of flexibility for the delivery of packages (such as emergency supplies in times of crisis).
Traditional UAVs and piloted aircraft employ a fixed-wing design composed of highly rigid materials which restrict landings to large prescribed areas and require long stretches of runway. In order to widen the window of possible landing sites, engineers at the University of Bristol and BMT Defence Services (BMT) have developed a morphed wing design that mimics the structure and operation of a bird’s wings in flight.
A legion of machine learning algorithms were recruited in order to control this new and more complex wing design, as they are able to quickly and efficiently articulate the series of landing movements which would be difficult for a human to master.
The UAV prototype passed the altitude testing phase, allowing researchers to move to the next phase of their work: completing repeatable ground landings. Tom Richardson, professor in the department of aerospace engineering at the University of Bristol, commented that, “The application of these new machine learning methods to nonlinear flight dynamics and control will allow us to create highly manoeuvrable and agile unmanned vehicles. I am really excited about the potential safety and operational performance benefits that these new methods offer.”
This breakthrough in autonomous landing capabilities is one aspect of an 18-month research project undertaken by the Defence Science and Technology Laboratory's (dstl) Autonomous Systems Underpinning Research (ASUR) program.
For more news about UAVs and machine learning algorithms, check out DARPA's UAV record for endurance and this new intelligent system for traffic management.