期刊名称：Journal of Theoretical and Applied Information Technology
印刷版ISSN：1992-8645
电子版ISSN：1817-3195
出版年度：2014
卷号：59
期号：3
出版社：Journal of Theoretical and Applied
摘要：The modeling of moving objects can attract the lots of research interests. The moving objects have been developed as a specific research area of Geographic Information Systems (GIS). The Vehicle movement location prediction based on their spatial and temporal mining is an important task in many applications. Several types of technique have been used for performing the vehicle movement prediction process. In such a works, there is a lack of analysis in predicting the vehicles location in current as well as in future. Accordingly we present an algorithm previously for finding optimal path in moving vehicle using Genetic Algorithm (GA). In the previous technique there is no complete assurance that a genetic algorithm will find an optimum path. This method also still now needs improvement for optimal path selection due to fitness function restricted to prediction of complex path. To avoid this problem, in this paper a new moving vehicle location prediction algorithm is proposed. The proposed algorithm mainly comprises two techniques namely, Particle Swarm optimization Algorithm (PSO) and Feed Forward Back Propagation Neural Network (FFBNN). In this proposed technique, the vehicles frequent paths are collected by watching all the vehicles movement in a specific time period. Among the frequent paths, the vehicles optimal paths are calculated by the PSO algorithm. The selected optimal paths for each vehicle are used to train the FFBNN. The well trained FFBNN is then utilized to find the vehicle movement from the current location. By combining the PSO and FFBNN, the vehicles location is predicted more efficiently. The implementation result shows the strength of the proposed algorithm in predicting the vehicle�s future location from the current location. The performance of the new algorithm is evaluated by comparing the result with the GA and FFBNN. The comparison result demonstrates the proposed technique acquires more accurate vehicle location prediction ratio than the GA with FFBNN prediction ratio, in terms of accuracy.
