Data Management Of Indoor Moving Objects

Modern indoor spaces usually accommodate large populations of individuals. People spend large parts of their lives in indoor spaces such as office buildings, shopping malls, conference centers, airports, metro and other transport infrastructures, etc. With appro-priate indoor positioning, the locations of indoor moving objects can be determined and recorded. The management of indoor moving objects can be used as the foundations for a series of indoor location-based services, including indoor navigation, staff security, space planning, advertisement and pricing. As a result, the effective and efficient management of indoor moving objects is very significant.In the recent decades, sufficient research on outdoor moving objects has been con-ducted on different indexing methods and various query processing techniques. However, these techniques cannot be adopted in indoor scenarios directly, because of two distinct characters in indoor space. First, complex indoor topology makes the traditional modeling and trajectory representations for outdoor space inappropriate. Second, indoor positioning techniques usually cannot report the exact locations continuously like what GPS-enabled devices do in outdoor space, which result in high uncertainty.In this thesis, a comprehensive survey on the techniques for managing the outdoor moving objects is conducted. In order to cater the indoor scenarios, a graph model is proposed as the foundation of indoor modeling. The tracking method, versatile index-ing structures and different query processing methods for spatio-temporal range queries, probabilistic continuous range monitoring and probabilistic threshold k-Nearest-Neighbor (kNN) queries are proposed on the graph model.The contributions of this thesis are summarized as follows:1. A versatile graph model which can capture both the topology of the indoor space and the deployments of symbolic indoor positioning techniques is proposed.2. Based on the graph model, an indoor tracking method which can be used in both off-line and on-line scenarios is developed.3. A novel symbolic representation for historical indoor trajectories is proposed. In order to support spatio-temporal range queries over the symbolic trajectories, two novel indexing methods, i.e. RTR-tree and TP2R-tree, are developed.4. Based on the state classification, a hashing-based indexing scheme is proposed to index the current locations of indoor moving objects. A query-aware and incremen-tal query processing method for probabilistic range monitoring over indoor moving objects is also developed.5. An uncertain model for the inaccurate locations of indoor moving objects is defined, and an efficient query processing method for probabilistic threshold kNN queries is developed.A comprehensive research on the management of indoor moving objects is conducted in this thesis. A versatile graph model is proposed as the foundation of the problem. And various query processing methods and indexing structures are developed for indoor spatio-temporal queries, probabilistic range monitoring queries and probabilistic threshold kNN queries.