From Robotic Fish to Hummingbirds: Embodied Intelligence in Aquatic and Aerial Locomotio

In bioinspired locomotion, the superior performance and intelligent behaviors observed in animals and robots often emerge from the dynamic interactions among body structure, musculoskeletal joint mechanics, neural reflexes, and the physical properties of the surrounding environment. We consider such dynamic interactions to give rise to a form of embodied intelligence.  In this talk, I will present three examples to highlight such embodied intelligence in aquatic and aerial locomotion. First, I will describe how intelligent swimming behaviors in a modular robotic fish, such as tandem swimming and efficient station keeping, emerge from hydrodynamic pressure feedback and joint torque control. Second, I will illustrate how the inverted landing behaviors of bluebottle flies inspire control strategies for dynamic perching in small robotic fliers, enabling rapid and robust landings on surfaces with varied orientations and approach conditions. Lastly, I will describe how hummingbirds achieve rapid disturbance rejection and exceptional whole-body aerial stability. This is accomplished through wing-body joint-level stability that exploits aero-inertial dynamics - specifically, Flapping Counter-Torque (FCT) - allowing them to reject external disturbances within a single wingbeat, even in the absence of active control. We hypothesize that such capabilities arise from the self-stabilizing properties of their wing musculature or reflexes at the spinal level. Together, these examples underscore the importance of embodied intelligence in achieving robust, adaptive, and efficient locomotion in both natural and robotic systems.