Continuous Evaluation Of Monochromatic And Bichromatic Reverse Nearest Neighbors Query Over Uncertain Moving Objects

In recent years, with the development of wireless network communication technology and positioning technology (GPS), more and more wireless devices (such as mobile phone, GPS, etc.) have brought a proliferation. When the equipment is deployed to the moving objects, people can get the location data of the equipment. Such applications involve objects and query points that move unpredictably, where their locations are updated frequently. In this highly dynamic environment, results of spatial queries are no longer static, but are constantly changing. Often, due to limitation of measurement devices, environmental disturbance, or characteristics of applications, the location data obtained from the real world are uncertain.Reverse nearest neighbors query over uncertain data was appeared based on the above background. In this paper, the continuous query algorithm over monochromatic objects is designed first. The so-called monochromatic query is that both the query and points are the same kind, such as find the reverse nearest neighbor of me in the war. In this kind of query, a new index technology (virtual quadtree) is used. Based on the virtual quadtree, a close monitor-region is built by drawing the perpendicular bisector between the query and its nearest point, also the nearest point is considered as the candidate point. Then we just judge the candidate point is whether the reverse nearest neighbor of the query or not. At last when the query or candidate point changes its location, the close monitor-region is needed to be checked.Secondly, the continuous query algorithm over bichromatic is designed. The difference between monochromatic query and bichromatic query is that bichromatic query focus on different kind of objects while monochromatic query deal with the same kind of objects. For example, a soldier searches its nearest tank. The method is the same as monochromatic query. The only difference is that after finding the close monitor-region the candidate points are got from the region rather than from its nearest neighbors.At last some experiments are conducted to compare and analyze the solutions. The experimental results show that these solutions have good scalability on the number of moving objects, moreover, this paper analyze the factors that affect the performance of the algorithms, such as uncertainty region and threshold.