Semantic Modeling and Reasoning at Runtime for Autonomous Systems Engineering

Autonomous computing has two main objectives: (i) making systems aware about their state and their environment and (ii) making them able to automatically actualize their behavior according to any relevant internal or external event. This paper describes a novel and dynamic approach to assist software developers and designers in building autonomous systems. Our approach offers a dynamic platform to deploy software components and automatically control them at runtime according to a continuously updated knowledge about their state and their runtime environment. This knowledge is automatically maintained by semantic models that keep the system aware about its state and its environment. Autonomous systems designers can edit these models at runtime to inject extra knowledge without disrupting the execution of the deployed components. Moreover, this platform facilitates the implementation of autonomous self-properties by defining semantic rules that (i) reason on the captured knowledge about the system environment and state at one hand and (ii) apply self-adaptive actions on the system components on the other hand. Our platform, entitled Automatic and Semantic OSGi Control (ASOC), combines the high dynamicity of OSGi and the power of semantic inferences to offer better situation analysis and self-adaptation results in autonomous systems.