چكيده به لاتين
The proliferation of positioning devices such as smart phones, RFID tags and vehicle navigation systems and development in wireless technologies have resulted in an increasing growth in location-based services. Tracking of moving in different applications such as Intelligent Transportation Systems (ITS), Location-Based Services (LBS), tourism and location-based social networks has resulted in massive amounts of data. The exponential increase in the amount of such trajectory data caused communicational and storage problems and it is difficult to run spatio-temporal queries. Storing and processing this amount of data requires a great deal of resources and time which highlight the need of scalable and efficient methods to store data and answer queries.
Spatial-temporal data indexing techniques, query processing methods, rate and accuracy are important challenges in moving objects databases (MODB) for this amount of data. Also, maintaining the quality of data considering application requirements is another challenge.
In present research, a solution is based on three steps of preprocessing, indexing and query processing has been presented. In the preprocessing step, the map-matched methods are used for increase the accuracy and quality of input data. Compressing and reducing data is the main part of the preprocessing phase and the analysis of spatial-temporal data. By reducing trajectory data, one can overcome to some these problems. We used a road network based on clustering technique and a training set which are used to reduce network matched trajectory data. In the indexing step, according to amount of dataset, three combinatorial indexing methods for historical data (trajectory) and current position of moving objects are presented. For constructing and storing of spatial-temporal indices in the main memory, the disk and distributed environments are used respectively. Also, the effect of data reduction methods on indexing and processing of spatio-temporal query were evaluated. Due to the diversity of applications and queries in ITS and LBS, in addition to providing a new taxonomy for some of these queries, an approximate algorithm is presented for processing the aggregate or group nearest neighbor search query by using the proposed indices. Finally, a combinatorial method was suggested to predict the position of moving objects by using the proposed indexing method and semantic data.
