SitCom - Development Platform for Multimodal Perceptual Services

In this paper we introduce SitCom , a novel software utility for developing context-aware services, editing and deploying context situation models, and simulating perceptual components. SitCom , which stands for Situation Composer, is constructed as an open and extensible framework with rich graphical rendering capabilities, including 3D visualization. One of SitCom 's main goals is to simulate interactions among people and objects in various settings such as presentation rooms, meetings halls, social places, homes, or car cockpits. Moreover, SitCom captures and models such environmental context and delivers the abstracted information to context-aware user services. Contextual models are designed as pluggable modules. They can be configured to build hierarchies of contextual models and deployed and reused across applications, agent systems or cognitive architectures. SitCom allows plugging-in various multi-modal sensing components such as body trackers, face detectors, speech recognition engines, gesture recognizers, etc. Therefore, switching from simulated to real components with SitCom is seamless, typically involving no change in the contextual models and the services.