An Architecture for Dynamic Context Recognition in an Autonomous Driving Testing Environment

The Internet of Things (IoT) envisions billions of connected objects being able to sense, act and communicate with each other. This situation provides rich opportunities to dynamic context recognition mechanisms, which aim to make use of the connected sensors to recognize desired context information. However, there are still several challenges to solve for dynamic context recognition mechanisms to be implemented. One of those challenges is dynamic resource selection, the ability to dynamically select between different data sources to detect a specific type of information and to adapt this selection, when necessary. In this paper, we present an agent-based architecture to combine event-driven Service-Oriented Architecture (SOA) with a data streaming platform to enable dynamic context recognition. The architecture has been defined in the scope of a testing environment for autonomous driving. There, it is used to dynamically select data sources to provide information to autonomous vehicles and applications based on them.