Enabling end-user specification and debugging of complex events for location systems

While real-time location systems are gaining popularity as a platform for a variety of applications (e.g., asset tracking, safety and security, workflow monitoring and optimization), recent studies have shown that two critical barriers to adoption remain: 1) cost of location sensors, and 2) cost of tuning events in location-aware applications. In this dissertation, we present the Cascadia middleware for specification, detection, and management of location events. We also present a suite of end-user tools which, together with Cascadia, significantly reduce barriers to adoption of real-time location systems. The Cascadia system leverages recent advances in probabilistic data management to facilitate specification, detection, and management of complex events over sporadic and incomplete location data. In particular, Cascadia extends the Lahar system [212, 2731 to detect an important family of location events as patterns over inferred probabilistic database views representing entities' location traces. This approach enables use of less reliable location sensors (e.g., passive RFID) that are 10 times cheaper per tracked entity than state-of-the-art sensors. Clear semantics and a pattern-oriented query language also enable Cascadia to expose an intuitive API for querying and subscribing to location events. In addition to tools that define metadata, we present the Panoramic tool, which supports end-users in customizing and tuning their applications by directly specifying and debugging location events. Panoramic does not require users to write code, understand complex models, perform elaborate demonstrations, generate test traces, or blindly trust deterministic events. Instead, it allows end-users to specify and edit complex events with a visual language that embodies natural concepts of space and time. It also takes a novel approach to verification in which events are extracted from historical sensor traces and presented with intelligible, hierarchical visualizations. Our work with Cascadia and Panoramic is grounded in a comprehensive survey of events in location-aware computing as well as a detailed analysis of real location data from experiments with a building-scale RFID system called the RFID Ecosystem. We refine our designs through iterative laboratory studies and conduct a summative evaluation through additional laboratory studies and a longitudinal study with over 60 participants and 100s of location sensors in the RFID Ecosystem.