Adaptable Software for Dependable Cyber-Physical Systems

The mobile researcher, Amir Taherkordi, started his postdoctoral position at the University of Oslo (UiO) in December 2012, working broadly on software design, distribution and adaptation for embedded systems, sensor networks, and IoT systems. Amir is inv ited to carry out a 6-month research visit to Carnegie Mellon University (CMU) in Pittsburg, USA. The computer science school of CMU has a long-standing research effort on software architecture design and dynamic adaptation of software systems. Over past few years, they have particularly focused on Cyber-Physical Systems (CPS) from the software architecture viewpoint. CPSs are the next computing revolution with applications in medical systems, environmental control, safety and control, etc. CPS applicatio ns are primarily expected to be reliable, safe and secure (i.e., dependability). To meet these requirements, the CPS software system should support dynamic software adaptation. The special attributes of CPSs, such as high complexity and high degrees of au tomation and control make software adaptation a unique challenging issue in this domain. This project aims to address this from two complementary views: system-view and architecture-view. UiO has long experience in addressing the former for embedded syste ms, while CMU has been focusing on the latter for conventional software platforms over past decades. This joint project is aimed at proposing a generic adaptation software framework, including the software architectural model, for CPSs that promises a sol id and reusable software solution for dynamic adaptation of CPSs. Using this framework, the stakeholders in this domain (developers of CPSs in the industry and the academia) will be able to produce more dependable software for CPSs with less development e ffort.