Resilient Architecture Framework for Robotic Systems
Anton D. Hristozov, Eric T. Matson, John C. Gallagher, Marcus Rogers, Eric Dietz
- Year
- 2022
- Citations
- 5
Abstract
Robotic architectures in use today provide some mechanisms for evolution but fall short of their capabilities for dynamic reconfiguration and adaptation during run-time. Adding new modes of operation, components, and connections allows for supporting the evolution of a product throughout long periods. We analyze and discuss the possibilities of performing architecture modifications on the fly without rebooting the systems since a reboot may be undesirable or impossible. For this, we rely on strict component interfaces and a component manager subsystem that can make the transition from one architecture to another while the robot is operating. Our study focuses on modern systems such as ROS2 and PX4 and assumes a publish-subscribe mechanism for communication. The ability to modify the software architecture allows for adaptive solutions that work in different environments and are less susceptible to specific attacks. Switching to a safer architecture can limit some functionality but can increase security. Working with structurally defined interfaces, we set the possibility of using COTS components that comply with our requirements to build new architectures. This can make reuse and composition easier and improve time-to-market while guaranteeing reliability and safety.
Keywords
Related papers
Statistical Learning Theory
Yuhai Wu, Vladimir Vapnik
1999
Artificial intelligence: a modern approach
1995
Fractional Differential Equations
Igor Podlubný
2025
Applied Nonlinear Control
Jean-Jacques Slotine, Weiping Li
1991