Research On Location Querying And Prediction In Mobile Environment

With the development of information technology, information products are free withoutlimitation of place, time, method, and mode etc. This requirement promotes the emergenceand development of mobile computing applications. In a mobile computing environment,wireless networks, such as GSM, IS-95, IRIDIUM, can be used to exchange users' locationinformation in real time. The information is obtained with the help of localizers, such asGlobal Position System (GPS). Thus Location-Dependent Query (LDQ) is derived. Thequery may occur only at one time point or last for an interval. The answer to a LDQdepends on the location of the object asking the query, which is called querying object, andthe locations of the queried objects involved in the database, which is called moving object. Records are static in a traditional DBMS, which means that all attributes are kept in thesame values as they are explicitly updated last time, even though the attributes are changedactually. Then, in order to correctly indicate real situation, moving objects have tocontinuously send their position information to the database through a wirelesscommunication channel. The costs of performance, wireless bandwidth and service areexpensive to update location information frequently. A LDQ that exists in the system for a period of time is called Location-DependentContinuous Query (LDCQ). In a given interval, a database server keeps sending queryingresults back to querying objects. A querying object may be a moving object involved inother queries. Movement of objects makes querying difficult to process. The answer to aLDCQ depends not only on the database contents but also on the time when the query isissued. Because the answer may be different from time to time, continuous evaluation isneeded. The answer to a LDCQ is presented as a set of tuples indicating that the object satisfies the query from begin time to end time. LDCQ isespecially useful for monitoring interested objects so that the querying object will be IVinformed immediately once they meet the querying conditions. The research on LDCQ is focused on two aspects: Moving Object Database (MOD)and location prediction updating strategy. This paper is based on the Moving ObjectsSpatio-Temporal (MOST) data model adopting dynamic attributes that is one of MODmodels. A dynamic attribute continuously changes as a function of time without explicitupdating. A querying result not only depends on the database constants, but also depends onthe initial querying time. The existed location prediction updating strategies considered the location deviation ofmoving objects. Unfortunately, they didn't consider the error ranges that guarantee theaccuracy of a querying result. This paper firstly analyzes the possible deviation scale ofquerying result that gives an error measuring standard for result accuracy. The error can beexpressed in location or time. As long as the deviation of a querying result does not exceedthe given error limitation, it is considered to be accurate. Since the actual location and timefor the intersection of a moving object and a querying boundary are difficult to bedetermined in advance, the predicted intersecting point is used in querying strategies.Selecting proper system parameters is discussed in this paper that regards the errorlimitation. There are two schemes ensuring error limitation. In the first scheme, at thepredicted intersecting time, the distance from the location of a moving object to thepredicted intersecting location is required to meet the given error limitation, as well as thedistance from the actual location of the predicted intersecting point on the queryingboundary to the predicted intersecting location. In the second scheme, only the distancebetween the location of a moving object and the actual location of the predicted intersectingpoint on the querying boundary is required to meet the given e...