Monitoring mobile cyber-physical systems by means of a knowledge discovery cycle

Abstract

Mobile cyber-physical systems (MCPSs) such as cars, trains, aeroplanes or spaceships are widely disseminated in our today's life. MCPSs are embedded into a physical environment which is usually harsh and uncertain. Hence, monitoring MCPSs is essential to ensure reliability and to avoid critical damage. Monitoring MCPSs is a very interdisciplinary research area. The monitoring process presupposes knowledge about the MCPS, the belonging physical environment and the related system behaviour. Such knowledge has to be discovered continuously during runtime due to continuous changes like wear and tear. Our contribution is the definition of an abstract process cycle for monitoring MCPSs which is called the Knowledge Discovery Cycle (KDC). The KDC is based on key challenges for monitoring MCPSs and common characteristics of MCPSs. We particularly discuss key challenges and common characteristics by means of the International Space Station (ISS). Based on this, we examine five overall considerations for monitoring MCPSs. Further on, we identify processing steps for the KDC and associate existing concepts and technologies with the identified processing steps. This involves knowledge discovery in databases, knowledge discovery from data streams, as well as information flow processing. As a result, our work outlines open problems for monitoring MCPSs.