Dynamic System Model Generation for Online Fault Detection and Diagnosis of Robotic Systems

TL;DR

This paper introduces a method for real-time dynamic system model generation to improve fault detection and diagnosis in complex robotic systems, reducing reliance on static models and expert intervention.

Contribution

It presents a novel approach that actively generates system models at runtime, enabling effective fault diagnosis adaptable to various robotic systems.

Findings

Enables real-time fault detection in robotic systems

Reduces dependence on pre-existing models and data

Minimizes computational overhead

Abstract

With the rapid development of more complex robots, Fault Detection and Diagnosis (FDD) becomes increasingly harder. Especially the need for predetermined models and historic data is problematic because they do not encompass the dynamic and fast-changing nature of such systems. To this end, we propose a concept that actively generates a dynamic system model at runtime and utilizes it to locate root causes. The goal is to be applicable to all kinds of robotic systems that share a similar software design. Additionally, it should exhibit minimal overhead and enhance independence from expert attention.