(via University of Washington)
As our self-sufficient robot building capabilities continue to develop, research seems to take a turn inward in the design of autonomous devices that attempt to mimic the human ways of doing things. Take, for instance, the quadrocopter that learns from its mistakes and iteratively changes its path every subsequent flight to improve its trajectory, speed, and time. In the case of these flying drones, why not look at a species that actually has their own set of wings? This sentiment is at the heart of Yonathan Munk’s most recent experiment over at the University of Washington in Seattle. By studying the flight patterns and trajectory of a hawkmoth in a virtual environment, Munk plans to help strategize new ways to improve the flying ability of drones.
The experiment was done by tying a string around a moth and allowing it to sway side to side as if it were turning. Munk then placed a screen that projected a virtual forest filled with bright blue poles that appeared as tree-like objects to the moth. The moth’s movement essentially mimicked the movement of a gaming joystick, dodging trees left and right. The moth even attempted to land on the virtual trees by flailing its legs to avoid crashing into the bright blue beams. Munk then made the game harder by creating a virtual fog that challenged the moth’s flight strategy in a lower visibility condition. The results were rather surprising as to the moth’s ability to strategize its flight techniques ahead of time - not what one would expect from the little bugger. Munk found that when visibility conditions were poor, the moth would fly in wide circles and land on a nearby tree used as a reference point before moving on to the next tree. With slightly better visibility conditions, the moth veered in the direction of oncoming trees as they appeared in the distance amidst the fog. Finally, with no fog at all, the moth would ignore distant trees and make turns according to the number of nearby trees.
(via University of Washington)
Munk’s research not only shows that these little flying creatures do an amazing job at planning their flight strategy ahead of time, but presents more information that can be applicable to the optimization of automated drones. Surely, there are several other animals that have naturally perfected abilities, humans would find useful to study and make use of. The information is all out there, it just needs to be found and implemented into the creation of more biologically developed technology.
Cabe
