Meeting Abstract
Flying animals must stabilize against environment perturbations from a variety of different sources, including aerodynamic effects such as wind gusts and physical effects such as mechanical contact with the environment or other animals. Many methods of studying perturbation response in flying animals rely on a single, brief perturbation event resulting in small datasets and varying perturbation inputs. Here we use a vortex chamber to apply continuous aerodynamic pitch perturbations to freely-flying hawkmoths (Manduca sexta), recorded with three high-speed cameras. Perturbations were delivered at 3 distinct magnitudes plus a hovering control, magnitudes approximated continuous pitch rotation of approximately 500, 1000 and 1500 degrees/s . We hypothesized that moths would respond to pitch-up perturbation by shifting the average stroke position rearward (and vice-versa for pitch-down perturbations) with approximately symmetric magnitudes, and also change wing orientation in each half-stroke by opposite directions for pitch-up and pitch-down, as suggested by recent momentary collision pitch perturbation experiments. Video data were analyzed using a deep-learning neural network trained to identify 6 landmarks on the moth wings and body. Results generally support our hypotheses, with the average stroke position in particular shifting as expected in the low and medium magnitude perturbations. However, compensation for the largest magnitude perturbation was qualitatively and quantitatively distinct and may reflect a shift to behavioral rather than biomechanical compensation.
