Addressing Dynamic Contextual Adaptation with a Domain-Specific Language

Abstract

The increasing number of mobile devices and sensors equipped with wireless networking capabilities enable a new generation of pro-active applications. These applications make use of context to adapt their behaviour to better fit their current situation. To support unanticipated changes to application behaviour, mechanisms are needed to specify when and how to adapt an application during its runtime. Many dynamic platforms exist that achieve this to some extent, and that are built on general-purpose languages (GPLs). However, these approaches suffer from standard difficulties of GPLs relating to the lack of semantic expressiveness of their constructs. In this paper, we describe high-level declarative constructs that can be used to specify the adaptation of application behaviour to specific situations. The language is supported by a framework that enables the exchange and merge of behaviours on-the-fly. Our approach is evaluated against application scenarios in the domain of autonomous vehicles.