Perception is a central issue for many applications in ubiquitous computing. However, in current implementations, sensing and distributed perception is re-invented over and over again. Identifying commonalities and synthesizing a common model would be more effective both from an economic as well as scientific perspective. We have analyzed the properties of ubicomp perception systems and used their characteristics as a main input for the design of a perception model. We present a layered perception model, discriminating functionality on artifact, setting, and application level. On the artifact layer, data collection, perception, and recognition for the particular artifact is modelled and implemented. The setting layer deals with perception and recognition tasks for a tightly coupled group of artifacts. In the application layer, perception and application-specific recognition resides. The approach was evaluated by building applications, two of which are also reported in the paper. The wider applicability of the model was assessed by analyzing further applications and how they map onto the artifact-based approach.