Meeting Abstract
Gliding animals frequently traverse complex spatial habitats to perform ecologically relevant behaviors. However, unlike flapping flyers, gliders are constrained in their ability to negotiate obstacles and cover longer distances by a fixed energy budget (takeoff height) and a relatively simpler wing anatomy. In this context, it is unclear how gliders achieve collision-free flight on a day-to-day basis. We address this question by quantifying voluntary glides in a naturally behaving, wild population of the flying lizard Draco dussumieri, inhabiting the Agumbe rainforest nature preserve in India. We digitized 25 glide trajectories of varying distances in 3D and found that Dracos executed non-equilibrium glides involving a steeper takeoff, shallower mid-glide and a sharper pitch-up landing maneuver to cover longer glide distances. Furthermore, we digitized the location of all trees in the recording environment and found that for a given glide distance, Dracos selected a target tree with less surrounding clutter and jumped away from an obstacle in line with the target tree. During flight, Dracos navigated their cluttered environment using a vision-based obstacle-avoidance model with up to 0.5g turns. Finally, to land on the target tree, Dracos initiated deceleration consistent with a visual trigger model and progressively increased their braking while approaching the landing tree. In summary, we used a biologically relevant framework of the environment, sensory input and the biomechanical capability of Dracos to show how gliders execute collision-free flight in a natural habitat.
