Evaluation of Prioritized Deep System Identification on a Path Following Task

This paper revisits system identification and shows how new paradigms from machine learning can be used to improve it in the case of non-linear systems modeling noisy unbalanced dataset. We show that using importance sampling schemes provide a significant performance boost modeling, which is helpful predictive controller. The approach first evaluated on simulated data Unmanned Surface Vehicle (USV). Our consistently outperforms baseline approaches this Moreover we demonstrate benefits methodology control setting. use model (USV) Model Predictive Path Integral (MPPI) controller perform track following task. discuss influence parameters prioritized outperform standard methods. Finally, apply real know-how developed here.