Emergence of low-dimensional motor features during sensory guided flight in flies
Abstract
Insect flight is achieved by a machinery that acts as a mechanical oscillator, with its rhythm and motion patterns emerging from coupled neural and mechanical systems. This machinery embodies elements of flight control but also imposes significant constraints, both of which remain poorly understood. In flies, despite a wealth of wing kinematics data, we still do not know how sensory inputs shape flies muscle actuation patterns (neural control) that drive their wing machinery via steering muscles (muscle function), and in turn how the muscles generate steering torques (aerodynamics and body dynamics). We hypothesize that insect flight maneuvers emerge from a low-dimensional basis or feature set in combined body motion space (i.e., Body Motion Primitives, BMPs) and wing joint torque space (i.e., Wing Motor Primitives, WMPs), as a result of collective steering muscle action. To identify BMPs and WMPs and to unravel how they are recruited in closed-loop flight control tasks, we will combine advances in machine vision to generate a rich set of 3D wing kinematics in tethered and free flight, inverse dynamics, dimensionalityreduction analytics and robophysics. The proposal has three broad objectives- 1) Implement a data- and model-driven exploration of Wing Motor Primitives (WMPs) during yaw turns, 2) Reveal how sensory information maps to WMP recruitment for task-level flight control, 3) Identify how flight BMPs and WMPs emerge in free flight maneuvers. We are poised to examine flight control in small flies such as Drosophila from the top down as advances in neurogenetics enable unparalleled bottom-up mechanistic description of muscle function. Taken together, the aims of this proposal will deepen the scientific understanding of the mechanisms that underlie the generation and control of complex aerial maneuvers. Further, this work will lay the groundwork for the development of autonomous aerial systems capable of flexible control and rich flight repertoires, thereby enhancing flight autonomy.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Mar 14, 2024
- Source ID
- FA95502310705
Entities
People
- Jean-Michel Mongeau
Organizations
- Air Force Office of Scientific Research
- Pennsylvania State University
- United States Air Force