Font Size: a A A

Sensor Networks, Approximate Query Processing

Posted on:2006-03-21Degree:MasterType:Thesis
Country:ChinaCandidate:M Y ZhangFull Text:PDF
GTID:2208360155461443Subject:Computer application technology
Abstract/Summary:PDF Full Text Request
With the development of the embedded technology, sensor network composed of cheap sensor nodes, which can be deployed to work without any infrastructures, has become possible. Its potential to be used in many fields makes it become a more and more important topic for scientists to study. Sensor network can be used to monitor the environment or to track moving objects. The problem of how to process the sensor data so that the data can be collected in the same way as in traditional databases has become a new field in the research of database systems. And it is also very important for the development of the research field of the parallel and distributed data processing. However, we are still facing some new challenges such as the mobile nodes, low bandwidth and disconnection property of the sensor network. There are many assumptions and limitations in the research work on sensor networks nowadays, which are not practical.This paper mainly studies the technologies of routing protocols in Sensor Network, data compression, data caching and approximation query. In this paper a new GPSR routing protocol based on group is presented. In the protocol, nodes of the network are partitioned into different groups. Each node maintains some routing information and can rout the messages using geographical information. The usage of this protocol lightens the load on network. This paper brings forward a dynamic group cache mechanism, which fully considers data access frequency, validity and consistency of data. It reduces the network load and improves the response speed by reducing the distance of data transmission. This paper puts forward an approximated join based on set, which reduces cost of communication and improves the efficiency of the query. The data pattern that we provided in this paper is the key to the applications of the sensor network. A data compression algorithmbased on the pattern is also presented. In the end, we show the performance and the cost of the algorithms by experimentations.
Keywords/Search Tags:Sensor Network, Approximate Query, Route, Cache
PDF Full Text Request
Related items