A Querying Method Over Moving Objects In Probabilistic Database

A query on database can be classified into Precise Query and Imprecise Query. When querying the location information of moving objects from database, the actual location information often changes faster than the data stored in the database because the objects move continuously. So when getting result data from database, the database usually has outdated. This will result in non-consistent between the result got from database and the actual one. For this situation, we cannot get the precise value of the result. So we usually use Probabilistic Query, which will add a probability value to the result, to present the fidelity of the result. We can get this probability value by using relevant probability calculation method. In the calculation models at present, when new update information arrives, the old one stored in the database will be replaced, the database only stores the data updated for the latest time. When we calculate the probability value using probability calculation method, we cannot get the statistics result of the location information, because there is no history record stored in the database. In order to get the probability value of the result, the present method is using generic distribution, such as Gaussian distribution. But this will reduce the accuracy for a special object. Additionally, for some special precise query, such as querying the location information of time point in the past, will become impossible. This will restrain the range of the query.For this situation, a new data processing method is proposed. This method changes the data stored manner currently used. When the updating information arrives, it will be inserted into the information table as a new record, instead of replacing the old one. Quantities of history records can be reserved by this way. This will solve the problem of some special Precise Query, such as query the location information of objects at the time point in the past. Besides that, for Probabilistic Query, using the data record to get the location distribution is more reliable and accurate than using the generic ones.In this paper, we will first introduce the Moving Objects Querying System. In this part, we will give out the system model and the function of the components of the model. After that, we will describe the most important part-the Location Information System in detail. In this part, we mainly introduce the location information updating controlling strategy for moving objects and the calculation of the probability of the result of Probabilistic Query. Then, compared currently used model, propose a new way of processing data, and supervise the calculation of two kinds of typical query. In the end, we summarize the whole text.