Percorrer por data de Publicação, começado por "2025-10-30"
A mostrar 1 - 1 de 1
Resultados por página
Opções de ordenação
- Neural network-based model for bioinspired underwater propulsion systemsPublication . Abreu, Fábio Emanuel Fiqueli; Lopes, Luiz Carlos Guerreiro; Camacho, Emanuel António RodriguesResearchers have been studying animal propulsion mechanisms to develop agile, efficient, and versatile biomimetic robotic systems for marine science, industry, defense, and recreation applications. This project concerns the development of an experimental methodology to analyze the propulsive performance of a system inspired by fish caudal fin kinematics. The experimental rig features a 3D-printed fin that oscillates with the help of a servo motor, whose frequency and amplitude vary over time, creating multiple test conditions. In the rig, a load cell is used to measure the thrust force generated by the oscillation, and an electric current sensor is used to measure the current consumed by the servo when making the movements. The data extracted from these experiments are used to train a Kolmogorov–Arnold Network (KAN), which learns to predict the thrust and power outputs of the system based on input parameters such as fin amplitude, frequency, and angular velocity. The results obtained offer experimental evidence supporting the viability of using neural networks to model thrust and power in bio inspired systems, while providing valuable findings that may contribute to the design and improvement of bio-inspired underwater propulsion systems.