Research Of Compound Indexing Structure For Moving Objects Trajectories In Road Network

With the rapid development of wireless communication technology, such as WIFI,3G, GPS etc., and the popularity of various positioning automotive equipment and wireless mobile terminals, more and more mobile objects location-based services has been used widely. Locations of moving objects change constantly over time. Spatio-temporal database which store the location information of moving objects should process both temporal and space data of moving objects. Traditional database is design to manage the static data, so it can’t adapt to process frequently changing data. In moving object database, moving objects indexing technique is the basis and core content of the moving object database management. In this paper, after in-depth analysis of the road network structure and the existing technology of moving object database, an efficient moving object trajectory index structure ETI (Efficient moving objects Trajectories Indexing) on read network is proposed.There are two levels, i.e. the upper one and the lower one, in The ETI tetanic. The upper level use a hash table and a2DR-Tree which stores the road of the network to index the road network space. The road hash table stores road identification and its section segments sets, while the2DR-Tree leaves store each road section spatial range. A lower1DR-Tree is pointed by the pointer of each node point. Given the structure of road network is fixed, rarely changed, and static for a long time, the road hash table and2DR-Tree can be treated as a static one. The lower structure of the ETI is composed of three parts, a moving objects hash table, a dynamically linked lists and1DR-Tree forest. The1DR-Tree is used to store the information of time and moving objects’ trajectory. The identity and real-time information of the moving object are stored in the moving object hash, which connected to the dynamic linked list by a pointer. The trajectory information of the moving objects is stored in the history linked list according to the update time-sequence. After analyzing the movement of moving objects in the traffic network, taking different levels of traffic complexity into consideration, the cache memory LRU replacement algorithm is introduced in the hash structure of moving objects and road. By doing this, the index rate of road and moving objects in memory increased, the visiting times of peripheral storage declined, and the efficiency of the index structure was improved. All these proved that the ETI has a good performance in indexing efficiency.Finally, in this paper, the T. Brinkhoff moving objects data generator simulating produce experimental data of moving objects on real Oldenburg network were also used. The Cooperation of the performance of ETI and FNR-Tree indexing was conducted by measuring the times of pages (I/O) accessing. Results shows that ETI has a better performance on moving objects trajectory update, window query and history trajectory query.