Data-Driven Machine Learning Models for a Multi-Objective Flapping Fin Unmanned Underwater Vehicle Control System

Abstract

Flapping-fin unmanned underwater vehicle (UUV) propulsion systems provide high maneuverability for naval tasks such as surveillance and terrain exploration. Recent work has explored the use of time-series neural network surrogate models to predict thrust from vehicle design and fin kinematics. We develop a search-based inverse model that leverages a kinematics-to-thrust neural network model for control system design. Our inverse model finds a set of fin kinematics with the multi-objective goal of reaching a target thrust and creating a smooth kinematic transition between flapping cycles. We demonstrate how a control system integrating this inverse model can make online, cycle-to-cycle adjustments to prioritize different system objectives.

Power-Aware Inverse-Search Machine Learning for Low Resource Multi-Objective Unmanned Underwater Vehicle Control (Student Abstract)

“…The solution, an inversesearch method, is typically difficult to implement on aerial and underwater control systems (Zhou, Gómez-Hernández, and Li 2012;Hansen and Cordua 2017), because it often requires searches that will invoke a forward model multiple Figure 1: An integrated inverse model on the control system times. While this is usually too computationally intensive for real-time control, we previously successfully developed an inverse model for UUVs, since they have a more flexible time constraint compared to other aerial and underwater control systems (Lee et al 2023). This inverse search model is able to optimize for two cycle-by-cycle performance metrics: thrust generation and smoothness of transitioning kinematics.…”