Meeting Abstract
The evasive banked turn of a fly is among the most rapid flight maneuvers in nature, which it executes using small adjustments in its wingbeat pattern. It is suggested that, after open-loop turn initiation, flies control the bank dynamics using a PI controller based on sensory input from halteres; the yaw rotations are suggested not to be controlled throughout the maneuver, resulting in large sideslip at the turn’s end. We tested these notions, by replaying banked turns of fruit flies on a newly-developed bio-inspired flying robot. Like insects, the robot steers by adjusting the motion of its flapping wings, and autonomous flight is achieved using on-board auto-pilot and sensors, including a haltere-like gyroscope. The robot’s banked turns, controlled using a gyro-based PI-like controller, resembled those of fruit flies remarkably well, suggesting that fruit flies use a comparable controller based on haltere input. Yaw dynamics was also similar between the fruit flies and robot, whereby both rotated into the turn. This yaw movement reduced sideslip and might thus increase escape performance. Because the robot’s yaw control was turned off, the yaw movement must have been produced passively. Using an aerodynamic model of flapping flight, we showed that a translation-induced coupled yaw torque caused this yaw movement. Because many flying animals tend to produce banked turns using flapping wings, the use of this mechanism might be more common in nature.
