Вы находитесь на странице: 1из 9

Finding Similarity of Moving data Using Trajectory Clustering

Shishir Shrivastava1, Ajay kumar Akasapu2, Lokesh Sharma3


Shishir Shrivastava1 Ajaya Kumar Akasapu2 Lokesh Kumar Sharma3 Student RCET Bhilai1CG1, Reader RCET Bhilai CG2,3
{shishir.contact1, akasapu2,lksharmain3}@gmail.com

Abstract trajectory data are this days very often and perspective of very large number of application so we tried to do some work on this area. This paper presents the work of clustering on the Trajectory Data and finding hidden patterns. We implemented density based clustering algorithm for trajectory and apply our data set on it. We found some interesting patterns. Keywords Trajectory Data Mining, Mobility data,

which cellular antenna each cell phone was associated at each time moment, and it might be used in a similar way for the decision-making process. When raw trajectory data are used in order to extract additional information, as occurs in recent trajectory researches, only the physical properties (spatial and temporal dimensions) use to be focused on (LEE; HAN; WHANG, 2007; GTING et al., 2000; GTING; SCHNEIDER, 2005; LAUBE; IMFELD; WEIBEL, 2005; WOLFSON et al., 1998). These approaches, which work with raw trajectory data may miss interesting information, since the background geographic information of the trajectories is not considered. If we relate the trajectorys background geography, we are able to discover useful information associated with the trajectory, like for instance the places visited by the trajectory or even the transition between those places. In Figure 1.1 it is possible to see the same set of trajectories considering the background geography and without considering it. So, intuitively we may check that if we have a geographic background context we have more options in order to aggregate more meaningful information than just consider the trajectory as an isolated object in the space.

INTRODUCTION Thanks to current sensors and GPS technologies, large scale capture of the evolving position of individual mobile objects has become technically and economically feasible. This opens new perspectives for a large number of applications (from e.g. transportation and logistics to ecology and anthropology) built on the knowledge of movements of objects (SPACCAPIETRA et al., 2008). The data provided by these technologies are raw data and it becomes difficult to work in a better level of understanding. Usually these data are made up of simple spatio-temporal points, in other words, the points have a spatial location and a time associated. When we sort these simple points by its ascending time we have intuitively a trajectory. Examples of possible trajectories could be the monitoring of wild animals, birds, people, automobiles, airplanes, ships, a soccer player, etc. Trajectories can be unidimensional, represented as a set of points, (e.g. individual people, birds) or multi-dimensional, represented by a set of polygons or complex objects (e.g. hurricanes, tsunamis, etc). In this dissertation we consider trajectories as a list of points. For details about monitoring of two or more dimensional objects we suggest the reading of Gtings work (GTING; SCHNEIDER, 2005). Usually, raw data are stored by companies only for operational purpose and are not used to extract useful information for decision-making. These raw data are normally represented as a set of single spatio-temporal points, which is intuitively known as trajectory. Usually the points are in the form (x,y,t) or (x,y,z,t), where x, y and z are the spatial dimension and t is the temporal one (KUIJPERS; OTHMAN, 2007). Examples of companies that keep this kind of data are: insurance companies, vehicle tracker companies and mobile carriers. Mobile carriers, despite not using GPS devices in order to obtain the raw data, have a similar log of cell phones, like for instance, when a mobile changes from one cell (range of a cellular antenna) to another. Logs keep records about

definition, a trajectory is limited to a discrete number of ordered time-space points: Definition 2. Trajectory. It is a list f(t0; x0; y0); (t1; x1; y1); (t2; x2; y2); :::; (tN; xN; yN)g, with ti; xi; yi 2 < for i = 0; 1; 2; :::;N and t0 < t1 < t2 < ::: < tN, where t0 is the instant when the object starts the travel and tN is the instant when the travel terminates.

Figure 1.1: The same set of trajectories being considered with and without geographic background (ALVARES et al., 2007) By associating the background geography it is possible to work in a higher abstraction level, a semantic one, closer to the human understanding. This recent approach was introduced by Spaccapietra (SPACCAPIETRA et al., 2008). In that work he proposed two new concepts, stops and moves, used to conceptually describe a trajectory. A stop represents an important part of the trajectory and a move represents transitions between stops. Trajectories From the users viewpoint, the concept of trajectory is rooted in the evolving position of some object traveling in some space during a given time interval. Thus, a trajectory is by definition a spatio-temporal concept. The concept of traveling object implies that its movement is intended to fulfill a meaningful goal that requires traveling from one place to another. Traveling for achieving a goal takes a finite amount of time (and covers some distance in space), therefore trajectories are inherently defined by a time interval (SPACCAPIETRA et al., 2008). A continuous trajectory can be formally defined as: Definition 1. Continuous Trajectory. A continuous trajectory T is the graph of mapping

In definition 1, the trajectory is seen as a continuous list of points in the continuous time dimension, and in definition 2, the trajectory is discretized. Most existing works covering trajectories of moving objects only the spatial and temporal dimensions are considered. Some focus on spatial properties, others on the temporal ones, and only a few merge space and time. In general, the trajectory is as an isolated object, associated with no other spatial entity, unless, in some cases, with other trajectories. Gting (GTING et al., 2000; GTING; SCHNEIDER, 2005; BECKER; JENSEN; SU, 2007) presents a full approach over modeling moving object databases (MODs). A MOD is a database developed to store and provide useful operations to manipulate trajectories, supporting both geometric and temporal properties. These works cover the whole development of a MOD, from query languages, passing through generic spatio-temporal data types and operations for these data. As MODs store moving objects and/or moving regions, they have specific data types in order to represent such elements, like moving points, multipoints, lines or regions. Some MODs use vectors in order to represent the speed of an object in order to decrease the number of stored points. Another important element in DBMS query languages are the operators. A MOD with support for many different kinds of spatial and temporal operators use to have a richer DBMS query language, since more options are available to perform queries. Examples of common operators are intersection, which tests if some object intersects another; distance calculates the euclidean distance between two elements; or semantic operators like may and must, which indicate uncertainly of the location of a point in a polygon (WOLFSON et al., 1998) or either in a cube in the (x,y,t) coordinates system. New operators used in existing query languages are necessary to obtain information about moving objects, like for instance, When and where was the spread of fires larger than 500 km2? or How many taxis are at most 10 minutes from Johns house?. Additionally, some uncertainty query mechanisms can be provided, since the locations of moving objects may not be exact. Wolfson (WOLFSON et al., 1998) tries to approximate the location of a moving element storing the speed vector. So, in order to avoid frequent updates in the object position, a vector that simulates the movement of the object is stored. As update the vector is less constant than update the own object coordinates, that approach avoids a bottleneck in the database.

where I is named time domain of T. In other words, a continuous trajectory is a tuple (x,y,t) where t is the time and (x,y) are the spatial coordinates associated, where t is a strictly positive natural number. However, as important as delimiting the time interval of a trajectory is discretizing it, since it is needed to restrict it to a finite list of points in order to work computationally. So, in the following

However, it is a tradeoff between the object precision location and the database updates. As his option was to reduce the database updates, precision is lost in the objects coordinates, and the queries are not precise. Brakatsoulas (BRAKATSOULAS; PFOSER; TRYFONA, 2004) combines trajectory data with spatio-temporal information. The geographic place modeled is the city of Athens. In their work they present a spatial mining language (SML) to answer queries and provide useful information about states of their MOD. In Brakatsoulas (BRAKATSOULAS; PFOSER; TRYFONA, 2004) model, each trajectory is seen as a line that goes up the temporal axis in a 3-dimensional space-time coordinate system. Additionally, the queries represent a polyedron in that spatio temporal axis, and depending on how the trajectories cross this region, they may or not be in the query answer, as shown in Figure 2.1 (a). In the same figure, in (b), are presented some basic spatial relations found between trajectories. Another approach about trajectories is seen at Guting (GTING; ALMEIDA; DING, 2004), where trajectories are seen over networks. As most routes scenarios can be modeled by a graph (edges and nodes), a network was a useful way to see a trajectory. The goal of that work is to provide a comprehensive data model and query language for moving objects in networks, supporting the description and querying of complete histories of movement. The main characteristics are: The model has a explicit concept of network embedded in space. A moving object is described relative to the network rather than the embedding space. The user extends the network information using the facilities of the DBMS. The model supports static or moving objects only in relation to the network model. There exists a framework to represent and query objects moving freely in the space.

Figure 2.1: In (a) the relations between trajectories and the environment; and (b) relations between trajectories (BRAKATSOULAS; PFOSER; TRYFONA, 2004) Conceptual Trajectories Some definitions about trajectories over both physical and semantic perspective are presented in this section. In Mark (MARK et al., 1999) trajectories are related to a large period of life (lifelines) and are associated to individuals locations at regular or irregular temporal intervals. The objective was to find spatial clusters in the past or to determine past environmental exposures, like, for example, causes of diseases and exposure to toxic substances. The period of life ranges from years to decades, since chronic diseases use to have as cause an extended exposure to an injurious antecedent. Figure2.2 shows three lifelines for several decades. Thriault (THERIAULT et al., 2002) followed a similar way building a spatio-temporal database model for handling lifelines allowing statistic analysis of any pre-defined event. Each lifeline is related to some aspect of the individual life, for instance, professional, residential, household, and so on, and it is divided in episodes and events, as can be seen in Figure 2.3. So, it is possible to find out evolution (or decision) patterns. For example, if we discover that 70% of people which lived in a industrial region for at least 5 years, would have, in a posterior moment of their life, some kind of cancer.

The use of a network as a model can be interesting because most trajectories happen in a well defined place, like street, fly routes, etc. In this dissertation we do not focus on network models because it is not generic enough to enclose all possible scenarios of trajectories. For instance, in the scenario of bird migration, we dont have routes well defined, as well as in a hurricane trajectory and other climatic scenarios.

These works started giving semantics to trajectories some years ago. However, a more recent approach came out, as will be explained in the next section.

Trajectory-Specific Clustering Purely distance-based methods impose some limitations both at the expressiveness level, i.e. some notions of clusters cannot be modelled in a handy way, and at the performance level, i.e. some opportunities for improving performances cannot be taken. That is mainly due to the strong separation between the dissimilarity criterion and the clustering schema that uses it without exactly knowing its semantics usually only assuming the distance is a metric. For instance, any method based on some notion of centre or, more generally, representative of a cluster needs to compute it in a way that is coherent with and thus dependent on the distance function adopted. The most prominent example is the basic k-means algorithm, where the representative is usually computed as the object (possibly new, not yet present in the input data) that minimizes the average distance between itself and all the objects in the cluster. In other cases, the model of cluster requested is not based on any notion of distance at all or, at least, not a distance between whole trajectories. Experimental results This chapter presents the experiments performed to validate the proposed method. In these experiments we used trajectory data collected by Microsoft Corporation. A common need in analysing large quantities of data is to divide the data set into logically distinct groups, such that the objects in each group share some property that does not hold (or holds much less) for other objects. As such, clustering searches a global model of data, usually with the main focus on associating each object with a group (i.e. a cluster), even though in some cases we are interested (also) in understanding where clusters are located in the data space. In this section, we focus on the context of moving objects and, thus, on the trajectories that describe their movement. In this setting, clustering consists essentially in trying to outline groups of individuals that show similar behaviours. Moving in the same way at the same time is sometimes too restrictive to discover useful information, and thus the temporal constraint might be removed. In these cases, we could look for groups of objects that follow the same route (i.e. the temporally oriented spatial projection of a trajectory) but at any moment in time, thus formulating requests of the type Find groups of individuals moving along the same roads, (10.12) for example, boats following some common itinerary to cross a sea, or cars following the same paths from home to workplace and reverse, etc. The bottom part of Fig. 10.1, where trajectories are spatially projected on the XY plane, shows a simple example of that, and the result is a unique cluster of objects that follow the same path, even though at different times and speeds. Again, in the time-series literature, we can find some general methods that yield some similar result. One is the comparison of pairs of time series by allowing (dynamic) time warping (e.g. see [R. Snodgrass at all] for an efficient approximation), i.e. a non-linear

Figure 2.2: Lifelines of three people showing the cities where they lived (MARK et al., 1999) Trajectory Data Mining In the area of data mining, we have seen a similar development. Many data mining techniques such as frequent set and association rule mining, classification, prediction and clustering were first developed for typical alphanumerical business data. From the second half of the 1990s, these techniques were studied for temporal and spatial data and sometimes specific, techniques such as time-series analysis were introduced in the data mining field. Cluster analysis' is a class of statistical techniques that can be applied to data that exhibit natural groupings. Cluster analysis sorts through the raw data and groups them into clusters. A cluster is a group of relatively homogeneous cases or observations. Objects in a cluster are similar to each other. They are also dissimilar to objects outside the cluster, particularly objects in other clusters. Types of Clustering Hierarchical algorithms find successive clusters using previously established clusters. These algorithms usually are either agglomerative ("bottom-up") or divisive ("top-down"). Agglomerative algorithms begin with each element as a separate cluster and merge them into successively larger clusters. Divisive algorithms begin with the whole set and proceed to divide it into successively smaller clusters. Partitioned algorithms typically determine all clusters at once, but can also be used as divisive algorithms in the hierarchical clustering. Density-based clustering algorithms are devised to discover arbitrary-shaped clusters. In this approach, a cluster is regarded as a region in which the density of data objects exceeds a threshold. DBSCAN and OPTICS are two typical algorithms of this kind.

transformation of time, so that the order of appearance of the locations in the series is kept, but possibly compressing/ expanding the movement times. Another method, proposed in [Ester M. at all] and further studied in [J.F. Allen at all], consists in computing the distance as length of the least common sub-sequence (LCSS) of the two series, essentially formulated as an edit-distance problem We conclude by mentioning the existence of other time seriesbased approaches that define distances between features extracted from series, rather than comparing the series themselves. For instance, we could extract all pairs of consecutive values in each series (in our context, consecutive locations within each trajectory) and then simply count the number of pairs shared by the two series compared, as proposed in [Ester M. at all] ; or, as an alternative, we could extract a set of landmarks for each time series (i.e. local behaviours of the time series such as minima or maxima or, more specific to our context, changes of speed or direction) and compute the distance between the series by simply comparing their corresponding series of landmarks, as described in [K. Patroumpas at all] Clustering Parameters The clusters were formally defined in section 3, but it is difficult to formally evaluate the quality of a cluster. Many clustering algorithms are evaluated through a visual representation. In a similar way, we use a visual representation of the trajectory and present the clustering result in a visual representation as well. Obviously clusters are important in order to show the quality of the clustering method. However, these clusters should not be obviously delimited in order to evaluate the influence of the clustering parameters in their quality. As the requairment of our algorithm we need some basic parameters which are

Output Data

Clustering Experiments After selecting the limited subset of trajectory and spacifing required para meters we have performed DBSCAN clustering in our data To find a cluster, DBSCAN starts with an arbitrary point p and retrieves all points density reachable from p wrt. Eps and MinPts. If p is a core point, this procedure yields a cluster wrt. Eps and MinPts. If p is a border point, no points are density-reachable from p and DBSCAN visits the next point of the database. Since we use global values for Eps and MinPts, DBSCAN may merge two clusters according to definition 5 into one cluster, if two clusters of different density are close to each other. Let the distance between two sets of points S1 and S2 be defined as dist (S1, S2) = min {dist(p,q) | pS1, q S2}. Then, two sets of points having at least the density of the thinnest cluster will be separated from each other only if the distance between the two sets is larger than Eps. Consequently, a recursive call of DBSCAN may be necessary for the detected clusters with a higher value for MinPts. This is, however, no disadvantage because the recursive application of DBSCAN yields an elegant and very efficient basic

Inputs Data Set

algorithm. Furthermore, the recursive clustering of the points of a cluster is only necessary under conditions that can be easily detected.

Which notions of similarity and distance are best suited for a specific (distance based) clustering task? In particular, different settings may require different levels of strictness in comparing trajectories: from checking spatial and temporal coincidence (trajectories are similar if they visit the same places at the same times) to spatial-only coincidence (the order of visit might be important, but not the precise timings), similarity of relative motions (considering speed, direction, etc.) or simply similarity of general features (average speed, length, etc.). Which notions of cluster best model the concepts of group peculiar to trajectory data? The complex nature of trajectories can lead to concepts of groups and clusters related to the inner structure of the data, such as the movement information relative to specific sub-intervals of time or sub-regions of space or, in other cases, they can require to develop general models of the overall movement, such as probabilistic models. Which features best model the kind of events or characteristics for which we want to extract classical local patterns, such as frequent item sets, sequential patterns and association rules? A wide range of alternatives is possible, in principle, ranging from simple aggregate information (e.g. length of the trajectory) to spatial or spatiotemporal descriptions of the movement (e.g. set of visited places, or maneuvers like U-turns). And we also implemented distance based clustering algorithm to find out some specific patterns. In future we tried to improve unsupervised learning using genetic algorithm for the Trajectory data set. By doing so we try to improve the performance of density based algorithm. And also try to find some interesting pattern as driver behavior. References

Fig : Clusters representing same interest

Fig:- Clusters representing same interest Projected View As shown in fig we clearly identify that data set which are presented in one color set are belongs to one cluster and also represent the interest of a person in one place at one time and the fig given below is specify the number of dataset in a cluster. Conclusion and Future Work Data mining on Trajectory data, and in particular on trajectory data, is a largely unexplored area. In this thesis, we presented several classes of problems that so far have been studied very little or, in some cases, not at all. The problems considered have been organized along a classical taxonomy of data mining tasks inherited from standard contexts, which include clustering, classification and local pattern tasks. Finally, because of their pervasive presence in the spatiotemporal context, also some problems raised by the uncertainty in the data have been briefly discussed. The main problems and issues pointed out include the following:

T. Abraham and J.F. Roddick. Survey of spatio-temporal databases. Geoinformatica, 3(1):6199, 1999. J.F. Allen. Maintaining knowledge about temporal intervals. Communications ACM, 26(11):832843, 1983. V.T. de Almeida and R.H. Guting. Supporting uncertainty in moving objects in network databases. In Proceedings of the 13th International Symposium on Geographic Information Systems (GIS05), pp. 3140, ACM, 2005. V.T. Almeida, R.H. Guting, and T. Behr. Querying moving objects in secondo. In Proceedings of Mobile Data Management (MDM06), p. 47, 2006. P. Bakalov, M. Hadjieleftheriou, E.J. Keogh, and V.J. Tsotras. Efficient trajectory joins using symbolic representations. In Proceedings of Mobile Data Management (MDM05), pp. 86 93, 2005. P. Bakalov, M. Hadjieleftheriou, and V.J. Tsotras. Time Relaxed Spatiotemporal Trajectory Joins. In Proceedings of

the 13th Annual International Workshop on Geographic Information Systems (GIS05), pp. 182191, 2005. Y. Bedard. Visual modeling of spatial databases: Towards spatial pvl and uml. Geomatica, 53:169185, 1999. Y. Bedard, S. Larrivee, M.-J. Proulx, and M. Nadeau. Modeling Geospatial Databases with Plug-Ins for Visual Languages: A Pragmatic Approach and the Impacts of 16 years of Research and Experimentations on Perceptory. In Conceptual Modeling for Advanced Application Domains, Vol. 3289, pp. 1730. Springer, Berlin Heidelberg New York, 2004. J. Bochnak, M. Coste, and M. Roy. Geometrie Algebrique Reelle. Springer, Berlin Heidelberg New York, 1987. Boosting location-based services with a moving object database engine. In Proceedings of the 5th Workshop on Data Engineering for Wireless and Mobile Access (MobiDE06). K. Borges, C. Davis, and A. Laender. Omt-g: An objectoriented data model for geographic applications. GeoInformatica, 5:221260, 2001. J. Brodeur, Y. Bedard, and M.-J. Proulx. Modelling Geospatial Application Database Using Uml-Based Repositories Aligned with International Standards in Geomatics. In ACM, editor, Proceedings of the 8th International Symposium on Geographic Information Systems (GIS00), pp. 3946, 2000. E. Camossi, M. Bertolotto, E. Bertino, and G. Guerrini. A Multigranular Spatiotemporal Data Model. In Proceedings of the 11th International Symposium on Geographic Information Systems (GIS03), pp. 94101, ACM, 2003. E. Camossi, M. Bertolotto, and E. Bertino. A flexible approach to spatiotemporal multigranularity in an object data model. International Journal of Geographical Information Science, 20(5), 2006. L. Chen, T. Ozsu, and V. Oria. Robust and Fast Similarity Search for Moving Object Trajectories. In F. Ozcan (ed.), Proceedings of the International Conference on Management of Data (SIGMOD05), pp. 491502. ACM, 2005. H. Darwen. Valid Time and Transaction Time Proposals: Language Design Aspects. In Temporal Databases: Research and Practice, LNCS 1399, pp. 195210, 1998. C. Date, H. Darwen, and N. Lorentzos. Temporal Data and the Relational Model. Model, Morgan Kaufmann, 2003. S. Dieker and R.H. Guting. Plug and play with query algebras: secondo - a generic dbms development environment. In Proceedings of the International Symposium on Database Engineering & Applications (IDEAS 00), pp. 380392. IEEE Computer Society, 2000. M.J. Egenhofer. Approximations of geospatial lifelines. 2003.

M.J. Egenhofer and R.D. Franzosa. Point set topological relations. International Journal of Geographical Information Systems, 5:161174, 1991. B. El-Geresy and C. Jones. Five Questions to Answer in Time: A Critical Survey of Approaches to Modelling in SpatioTemporal GIS, Chap. 3. GIS and Geocomputation-Innovations in GIS 7. Taylor & Francis, 2000. M. Erwig, R.H. Guting, M. Schneider, and M. Vazirgiannis. Spatio-temporal data types: An approach to modeling and querying moving objects in databases. GeoInformatica, 3(3):269296, 1999. M. Erwig and M. Schneider. Developments in SpatioTemporal Query Languages. In Proceedings of 10th International Conference and Workshop on Database and Expert Systems Applications (DEXA99), pp. 441449, 1999. M. Erwig and M. Schneider. Spatio-temporal predicates. IEEE Transaction Knowledge Data Engeneering, 14(4):881901, 2002. L. Forlizzi, R.H. Guting, E. Nardelli, and M. Schneider. A Data Model and Data Structures for Moving Objects Databases. In Proceedings of the International Conference on Management of Data (SIGMOD00), pp. 319330, 2000. E. Frentzos, K. Gratsias, N. Pelekis, and Y. Theodoridis. Nearest Neighbor Search on Moving Object Trajectories. In Proceedings of 9th International Symposium on Advances in Spatial and Temporal Databases (SSTD01), Vol. 3633. Lecture Notes in Computer Science, pp. 328345. Springer, Berlin Heidelberg New York, 2005. F. Geerts. Moving Objects and their Equations of Motion. In Proceedings of the 1st International Symposium on Applications of Constraint Databases, volume 3074 of Lecture Notes in Computer Science, pp. 4152. Springer, Berlin Heidelberg New York, 2004. T. Griffiths, A. Fernandes, N. Paton, and R. Barr. The tripod spatio-historical data model. Data and Knowledge Engineering, 49:2365, 2004. S. Grumbach, M. Koubarakis, P. Rigaux, M. Scholl, and S. Skiadopoulos. Spatio-temporal Models and Languages: An Approach Based on Constraints, Chap. 5, pp. 177201, 2003. R.H. Guting, M.H. Bohlen, M. Erwig, C.S. Jensen, N.A. Lorentzos, M. Schneider, and M. Vazirgiannis. A foundation for representing and quering moving objects. ACM Transactions on Database System, 25(1):142, 2000. R. Guting, M. Bohlen, M. Erwig, C. Jensen, M. Schneider, N. Lorentzos, E. Nardelli, M. Schneider, and J. Viqueira. SpatioTemporal Models and Languages: An Approach Based on Data Types. In Spatio-Temporal Databases: The Chorochronos Approach, LNCS 2520, Chap. 4, pp. 117176, 2003.

M.A. Hammad, W.G. Aref, and A.K. Elmagarmid. Stream Window Join: Tracking Moving Objects in Sensor-Network Databases. In Proceedings of 15th International Conference on Scientific and Statistical Database Management (SSDBM03), pp. 7584, 2003. K. Hornsby and M.J. Egenhofer. Modeling moving objects over multiple granularities. Annual Mathematics Artificial Intelligence, 36(12):177194, 2002. B. Huang and C. Claramunt. Stoql: An ODMG-Based SpatioTemporal Object Model and Query Language. In Proceedings of the 10th International Symposium on Spatial Data Handling (SDH02), pp. 225237, 2002. ISO/IEC. Information Technology Database languages SQL Part 7: Temporal (SQL/Foundation). ISO/IEC 9075-2 Working Draft. ISO, 2001. ISO/TC211. Geographic Information and Temporal Schema. ISO 19108:2002. ISO, 2002. ISO/TC211. Geographic Information and Spatial Schema. ISO 19107:2003. ISO, 2003. I. Kakoudakis. The Tau Temporal Object Model, M.Sc. Thesis, Umist, 1996. P.C. Kanellakis, G.M. Kuper, and P. Revesz. Constraint query languages. Journal of Computer and System Sciences, 51:26 52, 1995. V. Khatri, S. Ram, and R. Snodgrass. Augmenting a Conceptual Model with Geospatiotemporal Annotations. IEEE Transactions on Knowledge and Data Engineering, 16:1324 1338, 2004. B. Kuijpers, J. Paredaens, and D.V. Gucht. Towards a theory of movie database queries. In Proceedings of the 7th International Workshop on Temporal Representation and Reasoning (TIME00), pp. 95102. IEEE Computer Society, 2000. G. Kuper, L. Libkin, and J. Paredaens. Constraint Databases. Springer, Berlin Heidelberg New York, 2000. S. Larrivee, Y. Bedard, and J. Pouliot. How to Enrich the Semantics of Geospatial Databases by Properly Expressing 3d Objects in a Conceptual Model. In Proceedings of the Workshops On The Move to Meaningful Internet Systems, number 3762 in LNCS. Springer, Berlin Heidelberg New York, 2005. J.A.C. Lema, L. Forlizzi, R.H. Guting, E. Nardelli, and M. Schneider. Algorithms for moving objects databases. The Computer Journal, 46(6):680712, 2003. M.F.Mokbel, X. Xiong,W.G. Aref, S.E. Hambrusch, S. Prabhakar, andM.A. Hammad. Place: A Query Processor for Handling Real-ime Spatio-Temporal Data Streams. In

Proceedings of 30th International Conference on Very Large Data Bases (VLDB04), pp. 13771380, 2004. M.F. Mokbel, X. Xiong, M.A. Hammad, and W.G. Aref. Continuous query processing of spatio-temporal data streams in place. GeoInformatica, 9(4):343365, 2005. D. Papadias, J. Zhang, N. Mamoulis, and Y. Tao. Query Processing in Spatial Network Databases. In Proceedings of 29th International Conference on Very Large Data Bases (VLDB03), pp. 802813, 2003. D. Papadias, Q. Shen, Y. Tao, and K. Mouratidis. Group nearest neighbor queries. In Proceedings of the 20th International Conference on Data Engineering (ICDE04), pp. 301312. IEEE Computer Society, 2004. J. Paredaens, G. Kuper, and L. Libkin, editors. Constraint databases. Springer, Berlin Heidelberg New York, 2000. C. Parent. A Framework for Characterizing Spatio-Temporal Data Models. In S.S. Y. Masunaga (ed.), Advances in Multimedia and Databases for the New Century, pp. 8997. World Scientific, Singapore, 2000. K. Patroumpas and T.K. Sellis. Managing Trajectories of Moving Objects as Data Streams. In J. Sander and M.A. Nascimento, editors, Proceedings of 2nd International Workshop on Spatio-Temporal Database Management (STDBM04), pp. 4148, 2004. N. Pelekis. STAU: A Spatio-Temporal Extension to ORACLE DBMS. Ph.D. Thesis, 2002. N. Pelekis, B. Theodoulidis, I. Kopanakis, and Y. Theodoridis. Literature review of spatiotemporal database models. Knowledge Engeneering Review, 19(3):235274, 2004. N. Pelekis, B. Theodoulidis, Y. Theodoridis, and I. Kopanakis. An Oracle data cartridge for moving objects, laboratory of information systems, department of informatics, university of piraeus, unipi-isl-tr-2005-01, 2005. http://isl.cs.unipi.gr/db/ publications.html. N. Pelekis, Y. Theodoridis, S. Vosinakis, and T. Panayiotopoulos. Hermes A Framework for Location-Based Data Management. In Proceedings of 10th International Conference on Extending Database Technology (EDBT06), pp. 11301134, 2006. D.J. Peuquet. Making space for time: Issues in space-time data representation. Geoinformatica, 5(1):1132, 2001. D. Pfoser. Indexing the trajectories of moving objects. IEEE Data Engeneering Bullettin, 25(2):39, 2002. D. Pfoser and C.S. Jensen. Capturing the uncertainty ofmoving-object representations. In R.H. Guting, D. Papadias, and F.H. Lochovsky, (eds.), Proceedings of the 6th International Symposium on Advances in Spatial Databases

(SSD99), Vol. 1651. Lecture Notes in Computer Science, pp. 111132. Springer, Berlin Heidelberg New York, 1999. D. Pfoser, C.S. Jensen, and Y. Theodoridis. Novel Approaches in Query Processing for Moving Object Trajectories. In Proceedings of 26th International Conference on Very Large Data Bases (VLDB00), pp. 395406, 2000. R. Price, N. Tryfona, and C. Jensen. Extended spatiotemporal uml:Motivations, requirements, and constructs. In Journal of Database Management, 11:1427, 2000. R. Price, N. Tryfona, and C. Jensen. Extending UML for Space- and Time-Dependent Applications. Idea Group Publishing, 2002. S. Ram, R. Snodgrass, V. Khatri, and Y. Hwang. DISTIL: A Design Support Environment for Conceptual Modeling of Spatio-temporal Requirements, pp. 7083. 2001. P. Rigaux, M. Scholla, L. Segoufin, and S. Grumbach. Building a constraintbased spatial database system: Model, languages, and implementation. Information Systems, 28:563 595, 2003. A.P. Sistla, O.Wolfson, S. Chamberlain, and S. Dao.Modeling and Querying Moving Objects. In Proceedings of the 13th International Conference on Data Engineering (ICDE97), pp. 422432. IEEE Computer Society, 1997. R. Snodgrass, M. Bohlen, C. Jensen, and N. Kline. Adding valid time to SQL/Temporal. ANSI X3H2-96-501r2, ISO/IEC JTC1/SC21/WG3 DBL MAD-146r2, 1996. R. Snodgrass, M. Bohlen, C. Jensen, and A. Steiner. Adding transaction time to SQL/ Temporal: Temporal change proposal. ANSI X3H2-96-152r, ISO-ANSI SQL/ISO/IECJTC1/ SC21/WG3 DBL MCI-143. ISO, 1996. R. Snodgrass, M. Bohlen, C. Jensen, and A. Steiner. Transitioning Temporal Support in tsql2 to sql3. In Temporal Databases: Research and Practice, LNCS 1399, pp. 150194, 1998. Spatio-Temporal Databases: The CHOROCHRONOS Approach, Vol. 2520 of Lecture Notes in Computer Science. Springer, Berlin Heidelberg New York, 2003. J. Su, H. Xu, and O.H. Ibarra. Moving Objects: Logical Relationships and Queries. In C.S. Jensen, M. Schneider, B. Seeger, and V.J. Tsotras, editors, Proceedings of 7th International Symposium on Advances in Spatial and Temporal Databases (SSTD01), volume 2121 of Lecture Notes in Computer Science, pp. 319. Springer, Berlin Heidelberg New York, 2001. Y. Theodoridis. Ten benchmark database queries for locationbased services. The Computer Journal, 46(6):713725, 2003.

G. Trajcevski, O.Wolfson, K. Hinrichs, and S. Chamberlain. Managing uncertainty in moving objects databases. ACM Transactions Database System, 29(3):463507, 2004. N. Tryfona and C. Jensen. Conceptual data modeling for spatiotemporal applications. GeoInformatica, 3:245268, 1999. N. Tryfona, R. Price, and C. Price. Spatiotemporal Conceptual Modeling., chapter 3, pp. 79 116. 2003. M. Vlachos, D. Gunopulos, and G. Kollios. Discovering Similar Multidimensional Trajectories. In Proceedings of the 18th International Conference on Data Engineering (ICDE02), pp. 673684. IEEE Computer Society, 2002. N.V. de Weghe, F. Witlox, A.G. Cohn, T. Neutens, and P.D. Maeyer. Efficient storage of interactions between multiple moving point objects. In OTM Workshops (2), pp. 16361647, 2006. O. Wolfson, B. Xu, S. Chamberlain, and L. Jiang. Moving Objects Databases: Issues and Solutions. In Proceedings of the 10th International Conference on Scientific and Statistical Database Management (SSDBM98), pp. 111122, IEEE Computer Society, 1998. O. Wolfson, A.P. Sistla, S. Chamberlain, and Y. Yesha. Updating and querying databases that track mobile units. Distributed and Parallel Databases, 7(3):257387, 1999. J. Zhang and M. Goodchild. Uncertainty in Geographical Information. Taylor & Francis,2002 E. Zimanyi, C. Parent, and S. Spaccapietra. Conceptual Modeling for Traditional and Spatio-Temporal Applications The MADS Approach. Springer, Berlin Heidelberg New York, 2006. Ester M., Kriegel H.-P., and Xu X. 1995. A Database Interface for Clustering in Large Spatial Databases, Proc. 1st Int. Conf. on Knowledge Discovery and Data Mining, Montreal, Canada, 1995, AAAI Press, 1995.