Research On Location Based Mobile Information Service

Location-based services(LBS) are the delivery of data and information services where the content of those services is tailored to the current or some projected location.This is a new and fast-growing technology sector incorporating wireless technologies,mobile intelligent devices and Bei Dou positioning systems in China. In big data times,every person is both the producer and the consumer, which makes efficient management and intelligent exploration of massive data is the key mission of mobile services. The tasks for building a location based mobile service(LBMS) are essentially to collect, process, store, index, query, share and use location related massive data in a efficient and intelligent form.History location data(mainly GPS trajectories)is valuable for forecasting personal and social behaviors. So storage and use GPS trajectories is a important research scope of LBMS. The GPS trajectory database will be the basis of many intelligent applications,and need to support multiple user queries. The most common method of retrieving trajectory data is the spatio-temporal range query, such as: find all sub-trajectory within a given area and during a given time interval. However, temporal and spatial factors are not considered together in current history trajectory indices, resulting in the performance of answering the spatio-temporal queries is low. At the same time, the performance of indexes is sensitive to query size, can’t meet the needs of various query sizes.To enhance trajectory storage and index performance, this paper realizes Adaptive oc Tree based Trajectory clustering Index(ATTI): 1. Introduce trajectory storage distribution and time into the spatio-temporal range query cost modules. 2. Based on GPS update workload, dynamically adjustment the distribution of spatial division; 3. Using virtual octree forest implemented the query workload adaptive mechanism; 4. Base on real dataset of Microsoft Geo Life project do a wide range of experiments. The results show that compared with the multilevel index, ATTI can reduce half of query process delay.Location-based information sharing, as a important service of LBMS, is the base of many high level applications. In existing systems, the moving client needs to update its location on the server no matter whether any information needs to be shared, which makes the server a bottleneck for severe communication and processing overhead in large scale location based information sharing system. Our observation and experiment show that great overhead comes from spatial redundancy and information redundancy in the system. In this paper, Grid-based Indexing Mechanism(GIM)and Class-based Information Selection Mechanism(CISM) are proposed for on-demand information serving.In GIM, an information matrix is built in the server and synchronized to clients to indicate whether information implanted in the associated zone needs to be distributed to those clients. In CISM, on-demand information sharing will be implemented through clients’ customization on their interested information. The experimental results show that the scheme can eliminate about 70% communication overhead and work well especially for applications with unevenly distributed information.Protecting the Location privacy of users is requisite condition for popularizing a LBMS. At present, Location based mobile service providers, which are different companies or organizations, play the role of spatial data maintainers and spatial query processors. In order to handle privacy protected query, providers should implement specific query processors in their databases for different privacy preserving methods, which is infeasible for the diversity of providers and privacy preserving methods. What’s more,existing privacy preserving methods address location privacy threatens by revealing only the cloaked area or fake locations to server. It must induce some errors into query results and additional cost for dealing with cloaked spatial regions rather than exact location information. In this paper, we present PPSQP(privacy protected spatial query processor)to avoid the errors for k NN queries and range queries. A privacy protected k NN query is translated into one common k NN query and one range query without privacy decline and specific processor in the server side. Theory analysis and experimental result present PPSQP achieves location privacy protection with linear cost.Intelligence is the future of mobile information service. So, we introduce the methods of how to build a intelligent itineraries recommendation service make use of social networks and trajectory data as a example of LBMS. Human itineraries are often initiated by some general intentions, and will be optimized after considering all kinds of constraints and available information. In this paper, we propose a category-based itinerary recommendation framework to help user transfer intentions to itinerary planning, which joint physical trajectories and information of location based social networks. The main contributions are: 1. Build the category based activity scheduling model; 2. Design and implement the category tree based POI query strategy and algorithm; 3. Propose the Voronoi graph based GPS trajectory analysis method to build traffic information networks; 4. Combine social networks with traffic information networks to implement category based recommendation by ant colony algorithm. Experimental results show that, the recommendation results can be computed out online and have important reference value.LBMSs are also very useful in military scopes. So, in the last of the paper, we introduce how to build location based military services using above techniques in battlefield reconnaissance system, decision-making support system and logistics support system.