| Context Aware Computing:As a new computing form, context aware computing is closely related to pervasive computing, mobile computing and intelligent computing. Generally speaking, research frontiers such as pervasive computing, collaborative computing, intelligent computing and virtual reality are all context aware. This makes context aware computing a promising research area with vast applied prospects. With the developments of the Internet of things, sensor network, embedded system, communication network, distributed and mobile computing, context aware computing has drawn much attention from researchers both domestic and abroad. Meanwhile, more and more context aware applications have been developed. Through endowing the system the ability of sensing its surrounding environment, context aware computing optimizes the whole system (adjusting self behaviors, improving resource utilities), simplifies the interactions between the system and its users, improves users’experiences, makes the system intelligent, easy to use and user friendly.Context and spatial-temporal data:A huge amount of dynamic, heterogeneous, erratic and distributed data is generated and accumulated in context aware applications. Since time and space are the two basic components of these data, the achievements on spatial temporal data can be used in the research of these data. The researches on spatial temporal data aims to answer questions like "What has happened on it (him)? When, where and how does it happen?" These are quite similar with the4W (Who, When, Where and What) in context aware computing. Based on this reason, the researches on the huge amounts of data in context aware applications can be conducted from the spatial temporal characteristic perspective and combing the achievements on spatial temporal data research.The life circle of context:From the perspective of context life circle, context aware computing mainly involves context acquirement, management and usage. The context is acquired through analyzing user interfaces, logs, data semantic, physical and logic sensors. Storage and processing are the main parts of context management. In context storage, proper methods are used in representing, storing and indexing context. The integration of intelligent information processing technology and context aware computing provides more accurate and higher level information. Context usage is consisted of context query, analysis and etc. In this paper, researches on context awareness oriented spatial temporal data management, query processing, analysis and the related algorithms are conducted. The achievements will largely effect the application of context aware computing.Research scope:Although some progresses have been made on context aware computing and spatial temporal data, the new emerging context sensing devices, the emphasizing of user privacy, the higher requirement on query performance, and new emerging application scenarios make context aware computing a more challenging task. The research contents are summarized as follows:(1) The research statuses of context aware computing and spatial temporal data are surveyed. And through reviewing the history of context aware computing and spatial temporal data, existing problems are elaborated.(Chapter One)(2) Researches on context aware framework are conducted for effective acquiring contexts, masking heterogeneous data sources, improving the scalability, and protecting users’privacy. A prototype system is implemented to validate the proposed framework, it is also used in the implementations, validations and exhibitions of subsequent proposed methods.(Chapter Two)(3) Through the analysis of the characteristics of spatial temporal data, and the combination of existing achievements on temporal, spatial, relational and No SQL databases, a spatial temporal data management solution for context aware applications is designed. Based on the query requirements analysis of context aware applications, a spatial temporal index for context aware computing is proposed and validated to improve the query performance.(Chapter Three)(4) Since multiple location update strategies are existed, the sampled trajectories of one single object are not identical. A matching algorithm is required to distinguish the different samples of one trajectory from another trajectory. A dynamic time warping based trajectory matching algorithm is proposed. The proposed method can effectively attenuate the side effect of various location update strategies and improve date accuracy.(Chapter Four)(5) Since the rendezvous is a basic requirement in context aware computing, a thorough study of Group Nearest Neighbor query (GNN) is conducted. For emerging application scenarios, the Range based Probabilistic Group Nearest Neighbor query (RP-GNN) algorithm is proposed. Extensive experiments are conducted to validate the effectiveness, efficiency and scalability of the algorithm.(Chapter Five)(6) For the huge amount of data in some specified application domain (Such as public health. In this paper, searches are conducted based on the reasch project named Prediction System of Meteorology Related Diseases for Shanghai), analysis are conducted on the relationship among multiple context factors to discover the pattern and knowledge implied in the huge amount of data. Besides spatial and temporal characteristics, meteorological and environmental factors are also considered. The relationship between meteorological factors and mosquito density&activity is quantified. Meanwhile, the proposed context aware framework, data storage solution and methods on data acquiring, storing, processing and analysis are implemented and validated.(Chapter Six)The major contributions of our researches are summarized as follows:(1) The OWS-MA context aware framework is proposed and validated through the implementation of associated prototype system.(2) The design of data storage solution as well as the query performance improvement through the proposed HSTI spatial temporal index.(3) The Dynamic Time Warping base trajectory matching algorithm (DTW) is proposed. Its efficiency is validated through extensive experiments. Comparing with IMHD and OWD, DTW is also time sequence sensitive and time scaling tolerating.(4) The Range based Probabilistic Group Nearest Neighbor query (RP-GNN) algorithm is proposed. Its efficiency is validated through extensive experiments. RP-GNN reduces the computation complexity, speeds up the query processing, and acquires a better scalability.(5) The specific application (mosquito-borne infectious diseases prevention and control) is taken as a case study to extend the spatial temporal application into a multivariate data analysis. Through the comprehensive analysis of the relationship among multiple context factors, the relationship between meteorological factors and mosquito density&activity is quantified. Meanwhile, the proposed context aware framework, data storage solution and methods on data acquiring, storing, processing, and analysis are implemented and validated. The extension of spatial temporal application into a multivariate data analysis is feasible and effective. It is also an informatization breakthrough for mosquito-borne infectious diseases prevention and control... |
PDF Full Text Request