A gyroscopic control loop for active target vision
Abstract
Flight control in insects as well as human engineered vehicles makes use of motion cues both for target tracking and for balance. In animals including humans, these two sensory modalities split the frequency bandwidth of yaw rate control; mechanosensory equilibrium is high-pass (high frequency selective), whereas visual guidance is low-pass (low frequency selective). Therefore, in control systems the two modalities comprise separate, parallel, ‘inner’ (rapid) and ‘outer’ (slower) feedback loops. Keeping the two sensor pathways separate abrogates the need to resolve divergent timescales over which they operate, and instills control stability. The PI s present new evidence that visual processing is directly modulated by proprioceptive mechanosensory signals for active target vision. In a virtual reality flight simulator, rigidly tethered flies show classical high gain smooth motor responses both to stabilize panoramic motion and to pursue moving targets; flies that are magnetically tethered, free to maneuver in yaw, also show high gain smooth steering responses to panoramic motion, but unexpectedly nearly none to target movement. The PI s want to formalize in control theoretic terms the behavioral algorithms and functional neural circuits for these interactions. Discovering this new adaptive control topology could broadly impact the fields of sensory neurobiology and autonomous vehicle design. They suspect that short-delay high frequency proprioceptive feedback actively damps smooth optomotor steering responses to objects. The rationale for this hypothesis is that flies in a body-fixed flight simulator with disabled proprioception smoothly pursue a target moving in yaw, but in a body-free simulator with intact proprioception they do not. Using the formal language of system identification and control theory combined with open- and closed-loop control topologies, they will formalize mechanisms of visuo-mechanosensory integration during target-ground visual discrimination.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Mar 06, 2024
- Source ID
- FA95502310401
Entities
People
- Mark . Frye
Organizations
- Air Force Office of Scientific Research
- United States Air Force
- University of California, Los Angeles