2010•Journal of the Korea Society of Computer and InformationOpen access

An Indexing Scheme for Predicting Future-time Positions of Moving Objects with Frequently Varying Velocities

Sung-Chae Lim

Open full text 0 citations

Abstract

정보 기술과 이동 통신망의 발전에 따라 이동 객체의 위치 추적에 필요한 효과적인 색인 기법과 이를 통한 다양한 응용 서비스에 대한 사용자 요구가 커지고 있다. 이에 따라, 이동 객체의 현재 위치 추적과 미래 위치 추정에 사용되는 $TPR^*$ -트리와 같은 색인 기법이 관심을 받고 있다. $TPR^*$ -트리는 정적인 물체를 색인하기 위해 고안된 R-트리를 기본 구조로 하기 때문에, 갱신 비용이 크다는 단점이 있다. 따라서 이동 객체가 자주 속도나 위치 정보를 갱신하는 경우 트리 유지비용이 빠르게 증가할 수 있다. 특히, 이동 객체가 빠른 속도로 이동하는 경우 넓은 범위에 걸쳐 불필요한 노드 갱신이 발생할 수 있다는 문제점이 있다. 이런 문제점을 피하기 위해 논문에서는 루트노드의 자식노드에 이런 빠른 이동 속도의 객체만을 따로 색인하도록 하여 노드 갱신 비용을 크게 줄이는 방법을 제안한다. 제안된 방법을 통해 노드 갱신 비용을 최소화 하면서도 $TPR^*$ -트리의 장점을 유지할 수 있었다. 이런 성능상의 장점을 보이기 위해 시뮬레이션 기법을 사용한 성능 비교를 수행하였다. With the advances in the information technology and mobile communications, we now face increasing demands for various services based on both of position tracking of moving objects and their efficient index scheme. Accordingly, the $TPR^*$ -tree, which were proposed for efficiently tracking moving objects and predicting their positions in the future time, has drawn much intention. As the $TPR^*$ -tree came from the R-tree that is suitable for indexing static objects, it does not support cheap update costs. Therefore, it seems to be very costly to index moving objects if there are frequent occurrences of node updates caused by continuously changing velocities and positions. If some moving objects with high velocities have node updates, in particular, then the $TPR^*$ -tree may suffer from many unnecessary updates in the wide range of tree regions. To avoid such a problem, we propose a method that can keep fast-moving objects in the child nodes of the root node, thereby saving node update costs in the $TPR^*$ -tree. To show our performance advantages and retaining $TPR^*$ -tree features, we performed some performance experiments using a simulation technique.

Open-access reader

About this research paper

What this paper is about

정보 기술과 이동 통신망의 발전에 따라 이동 객체의 위치 추적에 필요한 효과적인 색인 기법과 이를 통한 다양한 응용 서비스에 대한 사용자 요구가 커지고 있다. 이에 따라, 이동 객체의 현재 위치 추적과 미래 위치 추정에 사용되는 $TPR^*$ -트리와 같은 색인 기법이 관심을 받고 있다. $TPR^*$ -트리는 정적인 물체를 색인하기 위해 고안된 R-트리를 기본 구조로 하기 때문에, 갱신 비용이 크다는 단점이 있다. 따라서 이동 객체가 자주 속도나 위치 정보를 갱신하는 경우 트리 유지비용이 빠르게 증가할 수 있다. 특히, 이동 객체가 빠른 속도로 이동하는 경우 넓은 범위에 걸쳐 불필요한 노드 갱신이 발생할 수 있다는 문제점이 있다. 이런 문제점을 피하기 위해 논문에서는 루트노드의 자식노드에 이런 빠른 이동 속도의 객체만을 따로 색인하도록 하여 노드 갱신 비용을 크게 줄이는 방법을 제안한다. 제안된 방법을 통해 노드 갱신 비용을 최소화 하면서도 $TPR^*$ -트리의 장점을 유지할 수 있었다. 이런 성능상의 장점을 보이기 위해 시뮬레이션 기법을 사용한 성능 비교를 수행하였다. With the advances in the information technology and mobile communications, we now face increasing demands for various services based on both of position tracking of moving objects and their efficient index scheme. Accordingly, the $TPR^*$ -tree, which were proposed for efficiently tracking moving objects and predicting their positions in the future time, has drawn much intention. As the $TPR^*$ -tree came from the R-tree that is suitable for indexing static objects, it does not support cheap update costs. Therefore, it seems to be very costly to index moving objects if there are frequent occurrences of node updates caused by continuously changing velocities and positions. If some moving objects with high velocities have node updates, in particular, then the $TPR^*$ -tree may suffer from many unnecessary updates in the wide range of tree regions. To avoid such a problem, we propose a method that can keep fast-moving objects in the child nodes of the root node, thereby saving node update costs in the $TPR^*$ -tree. To show our performance advantages and retaining $TPR^*$ -tree features, we performed some performance experiments using a simulation technique.

Why it matters

A significance statement is not available in the OpenAlex record.

Key contribution

A contribution statement is not available in the OpenAlex record.

Method / approach

Method details are not available in the OpenAlex metadata.

Main findings

Findings are not separately available in the OpenAlex metadata.

Limitations

Limitations are not available in the OpenAlex metadata.

Applications

Application details are not available in the OpenAlex metadata.

Available abstract

정보 기술과 이동 통신망의 발전에 따라 이동 객체의 위치 추적에 필요한 효과적인 색인 기법과 이를 통한 다양한 응용 서비스에 대한 사용자 요구가 커지고 있다. 이에 따라, 이동 객체의 현재 위치 추적과 미래 위치 추정에 사용되는 $TPR^*$ -트리와 같은 색인 기법이 관심을 받고 있다. $TPR^*$ -트리는 정적인 물체를 색인하기 위해 고안된 R-트리를 기본 구조로 하기 때문에, 갱신 비용이 크다는 단점이 있다. 따라서 이동 객체가 자주 속도나 위치 정보를 갱신하는 경우 트리 유지비용이 빠르게 증가할 수 있다. 특히, 이동 객체가 빠른 속도로 이동하는 경우 넓은 범위에 걸쳐 불필요한 노드 갱신이 발생할 수 있다는 문제점이 있다. 이런 문제점을 피하기 위해 논문에서는 루트노드의 자식노드에 이런 빠른 이동 속도의 객체만을 따로 색인하도록 하여 노드 갱신 비용을 크게 줄이는 방법을 제안한다. 제안된 방법을 통해 노드 갱신 비용을 최소화 하면서도 $TPR^*$ -트리의 장점을 유지할 수 있었다. 이런 성능상의 장점을 보이기 위해 시뮬레이션 기법을 사용한 성능 비교를 수행하였다. With the advances in the information technology and mobile communications, we now face increasing demands for various services based on both of position tracking of moving objects and their efficient index scheme. Accordingly, the $TPR^*$ -tree, which were proposed for efficiently tracking moving objects and predicting their positions in the future time, has drawn much intention. As the $TPR^*$ -tree came from the R-tree that is suitable for indexing static objects, it does not support cheap update costs. Therefore, it seems to be very costly to index moving objects if there are frequent occurrences of node updates caused by continuously changing velocities and positions. If some moving objects with high velocities have node updates, in particular, then the $TPR^*$ -tree may suffer from many unnecessary updates in the wide range of tree regions. To avoid such a problem, we propose a method that can keep fast-moving objects in the child nodes of the root node, thereby saving node update costs in the $TPR^*$ -tree. To show our performance advantages and retaining $TPR^*$ -tree features, we performed some performance experiments using a simulation technique.

Key concepts: Search engine indexing, Node (physics), Tree (set theory), Computer science, Data mining, Scheme (mathematics), Index (typography), Position (finance)

Related papers

Back to paper searchBrowse research topicsOriginal source
An Indexing Scheme for Predicting Future-time Positions of Moving Objects with Frequently Varying Velocities — Research Paper | ScholarLens