An integrated model-based diagnosis and repair architecture for ROS-based robot systems

Abstract

Autonomous robots are artifacts that comprise a significant number of heterogeneous hardware and software components and interact with dynamic environments. There fore, there is always a chance of faults at run-time that negatively affect the reliability of the system. In this paper we present a novel diagnosis and repair architecture for ROS-based robot systems. It is an extension to the existing ROS diagnostics stack and follows a model-based diagnosis and repair approach.In the paper we discuss the integrated diagnosis and repair architecture in detail. Moreover, we show its application to an example robot system and report first experimental results. The presented work provides three major contributions: a combi nation of diagnosis and repair, the integration of hardware and software, and the integration into ROS.

Formal verification of complex robotic systems on resource-constrained platforms

“…In [1], D-Finder, an invariant-based verification tool, is used to formally verify BIP (formal framework) models synthesized from functional robotic specifications. The verification of ROS [26] robotic specifications is the subject of several works [16,17,35]. We find in the literature also works using other techniques such as Discrete-Time Markov Chains models [10].…”

Autonomous skill-centric testing using deep learning

“…Even though this approach is similar in nature to our framework, it is restricted to and designed for low-level components. Zaman and Steinbauer et al describe a diagnosis framework and augment it with the additional ability of autonomous selfrepair [8], [9]. The framework is embedded into the ROS diagnostics stack in which single components can publish observations and diagnostic messages.…”

Autonomous fault detection for performance bugs in component-based robotic systems

“…A method that focuses on this part of robotic systems has been presented in different development stages by Zaman et al [2], Weber and Wotawa [23], and Kleiner et al [24]. The fault detection mechanism is based on observers which either check invariants in the communication or encode metrics as Horn clauses.…”

“…Detecting such faults allows to react to them in a proper way, potentially with a successful repair strategy [1], [2]. The need to detect and handle faults at runtime becomes even more important in robotics with the recent advances in physical human-robot interaction in industrial settings, where any fault can result in serious injuries, even at the development stage.…”