Our non-feathered friends: sensing and control in insect-scale flight

Abstract

Insects are capable of many tasks that insect-scale flying robots cannot yet perform, such as dodging waving grasses or branches, catching other insects as aerial predators, and taking off from and landing on a wide variety of surfaces. There has been significant progress in getting insect-scale flying robots off the ground in recent years, but to perform at the same level as insects, many advances will need to be made. These advances will come from both applying and adapting existing engineering techniques, and from interdisciplinary efforts towards gaining a better understanding of biological insects. Here, we present three ways to approach insect-scale flight. First, we develop a system identification method that is well suited to the unsteady nature, flexible wings, and viscous Reynolds number regime of insect-scale flight. The method builds a linear state-space model from a quasi-impulse response of an aeroelastic system. Second, we seek to better understand possible tradeoffs associated with aerodynamic force production and gyroscopic sensing in insects. In particular, we look at how wing and haltere shapes can affect the relative forces and strain associated with flapping and body rotation. Third, we provide a resource for bringing together animal movement and linear control theory. Feedback control is a fundamental concept in animal movement, and to translate advances in biological understandings to engineered vehicles, we must be able to describe and compare the dynamics of biomechanics and nervous systems into control architectures.