Eventos Anais de eventos
DINAME 2023
XIX International Symposium on Dynamic Problems of Mechanics
Physics Informed Machine Learning for Path Planning of Space Robots
Submission Author:
Rogerio Santos , SP
Co-Authors:
Rogerio Santos, Ijar DA FONSECA, Domingos Rade
Presenter: Rogerio Santos
doi://10.26678/ABCM.DINAME2023.DIN2023-0030
Abstract
The present study concerns a specific characteristic of the autonomy of space robots, namely the ability to plan the movement of the manipulators in the presence of uncertainties that affect the position of the target, normally due to the dynamic characteristics of the equipment. A physics informed machine learning strategy for the computation of the inverse kinematics of a space robot is proposed. The effects of satellite and manipulator dynamics are taken into account. The performance of the methodology is discussed as the amount of historical data increases. It is found that in some cases the machine learning algorithm is able to calculate the inverse kinematics with great precision. In general, its use results in a significant improvement on the process convergence.
Keywords
Support vector machine, Inverse Kinematics, uncertainty quantification, berthing, Space Robotics
