Optimization of robot structures and control schemes remains challenging despite advances in computing and rapid prototyping technology. Nature provides numerous examples of locomotion through multiple modes, such as crawling, swimming, and even jumping. However, the performance of state-of-the-art robots falls short of what is displayed by animal exemplars. A multi-modal computational paradigm would increase understanding of animal locomotion and allow improved robot designs.

The Pacific Lamprey is an example of an animal with unique swimming-climbing multi-modal locomotion. This talk presents a reduced order hybrid dynamic model for a lamprey-inspired wall climbing gait and hardware results for the performance of the lamprey-inspired wall climbing robot Trident. The hybrid dynamic model reveals insights for steering a lamprey-inspired climber and optimizing climbing efficiency.

Swimming robots pose a particular challenge for modeling due to the high computation cost of simulating fluid interaction. This talk presents a fluid interaction modeling paradigm that is computationally cheap, using theories by M.J. Lighthill and combining them with beam theory. This talk validates the model's predictive power by comparing it with hardware results. Further, the talk presents the model's utility by using it to co-optimize a control policy and tail morphology for the soft swimming-climbing robot SLIDER.

This talk will also apply these insights to the challenge of designing soft salp-inspired swimmers and soft robotics for coordinated manipulation in the marine environment.