| With the development of indoor positioning technologies, especially with the pratical use of new positioning devices like RFID, Wi-Fi, and Bluetooth, obtaining locations of indoor moving objects has become possible, and indoor location-based services(LBS) are developing rapidly. However, existing workes for outdoor spaces are not suitable for indoor environment because of the differences between indoor and outdoor spaces. These differences include space constraints, positioning methods, distance measurement, etc. Therefore, how to manage indoor moving objects efficiently to support indoor LBS has become a hot topic both in academia and in industries.Indoor navigation service is a major concern in indoor LBS. Generally, different people have different demands on indoor navigation service, so we have to consider many aspects when providing indoor navigation service. For example, we have to consider the complex structures of indoor spaces as well as various semantic constraints during navigation. This paper focuses on semantics-constrained navitgation in indoor space, which is a new and important issue in indoor LBS. Because existing researches on indoor navigation do not provide good solutions to this problem, this paper presents a new method to process indoor semantics-constrained navigation queries. Further, we develop a prototype system supporting indoor moving objects management that is also able to process indoor semantics-constrained navigation queries efficiently.The main contributions of this paper can be summarized as follows:(1) We study the indoor semantics-constrained navigation, and propose a navigation algorithm that aim at Include constraints and Exclude constraints. We analyze the application demands of semantics-constrained navigation, and define two constraints:Include constraints and Exclude constraints, which represent the semantic set that users want to pass and avoid respectively. The order traversal of semantic keywords in Include constraints always bring vast of repetitive computation. This paper proposes an algorithm that can avoid the order traversal. The algorithm divides paths into different categories according to the semantics of path, and then, all paths in each category has the same semantics and can be expanded concurrently using parallel Dijkstra. Exclude constraints makes the path must be real-time expanded. We take the indoor multilayer characteristics into account, and divide all floors into two categories, RTF and NRTF, which mean real-time floor and non-real-time floor respectively. And based on this, we propose a layer pruning optimization technique, which can prune large amounts of path expansion in NRTF. At the end, we discuss the continuous path update method of our query processing algorithm in real application scenario. The experimental results on simulation data demonstrate the effectiveness of our algorithom.(2) This paper designs and implements an indoor moving objects management prototype system call IndoorDB, which can support the semantics-constrained navigation service. IndoorDB extends and improves indoor space model LayeredModel. LayeredModel lacks of floor distribution and semantic information of indoor entity as well as the direction of doors, which are all needed for semantics-constrained navigation. And the difinition and storage of distance is not suitable for semantics-constrained navigation. Aiming at these problems, the paper extends the representation of room and door, defines a new indoor distace and storage mode, and proposes door queue and room queue handle door and room respectively. These extension and improvement can support the semantics-constrained navigation effectively. IndoorDB is implemented through extending indoor spatial-temporal data types and operations based on object-relational database system Oracle11g. IndoorDB is SQL-compatible. Thus it can be easily used in real database applications. |
PDF Full Text Request