On Complex Event Processing for Real-time Situational Awareness Nenad Stojanovic 1 , Alexander Artikis 2 1 FZI Karlsruhe, 76131 Karlsruhe, Germany Nenad.Stojanovic@fzi.de 2 NCSR Demokritos, Athens 15310, Greece a.artikis@iit.demokritos.gr Abstract. In this paper we give an overview of the existing research results and open research challenges in applying complex event processing for real-time situational awareness. We consider two different viewpoints: better detection of emerging complex situations and prediction of future situations. In order to illustrate these viewpoints we consider two application areas: activity recognition from the video content and social media observation, respectively. Keywords: Situation awareness, Social media observation, Activity recognition 1 Introduction Real-time data processing has become very important for many applications, such as item-tracking in RFID-supported logistics, social-media channels observation, activity recognition from video content, computer network monitoring, patient monitoring and trader behavior evaluation in financial markets. In all of these applications the amount of data being generated requires on–the-fly processing and immediate reaction in order to be managed in an efficient way. Indeed, such real-time orientation enables the detection of problems (e.g. a damaged item in a delivery, or bad image of a company in recently posted tweets) as soon as they happen, so that a corresponding reaction can be successfully performed. In the nutshell of this mechanism is the ability to recognize in real-time 1 (or even ahead of time) some interesting situations, what is called “real-time situational awareness”. Note that this goes beyond the traditional (static) situational awareness that is focused on the understanding a situation (if possible in real-time). Real-time situational awareness introduces the notion of real-time emergency: the main goal is to recognize a situation of interest as soon as possible in order to be able to react to it properly. Such a process introduces several challenges for the processing of data: 1. it should be very efficient in order to deal with a huge amount of events, 1 We consider „business real-time“ as the criteria for declaring something to be processed in real-time