System State Variable Discovery Counter Example

Abstract

The stability of an existing non-linear system model of a Dubins vehicle is investigated. The discovery of the system state variables and their dynamics from simulation data is attempted, using machine learning (ML) techniques. The results show that there is at least one case in which system state-space discovery through a ML approach is unsuccessful. Technology which relies upon state abstraction from inferential learning techniques may be vulnerable to failure if the cases are not well understood.