|
|
|
Estimating Missing Data in Sensor Network Databases Using Data Mining to Support Space Data Analysis
|
|
|
| Recent advances in Micro Electro Mechanical Systems (MEMS) based sensor technology, low-power analog and digital electronics, and low-power Radio Frequency (RF) design have made possible the development of relatively inexpensive and low-power wireless micro sensors that can be integrated in a network. The purpose of such a network is to monitor, combine, analyze and probably respond to the data collected by hundreds (or even thousands) sensors distributed in the physical world in a timely manner. This network can be used to support space data collection and analysis. For example, to facilitate solar system exploration missions, mobile sensors mounted on robots as well as hundreds of static micro sensors can be placed on MARS to collect its data and to send the collected data to a base station residing on MARS for real-time data analysis. The base station can then use the analysis results in real-time to determine actions that the robots should take next. However, in a wireless sensor network, a significant amount of sensor readings sent from the sensors to the data processing point(s) (servers) may be lost or corrupted. In this research we propose a power-aware technique that uses association rules mining to handle such a problem. In this technique, to save battery power on sensors and to meet real-time requirements for data analysis, instead of requesting the sensor nodes (MS), the readings of which are missing, to resend their last readings, an estimation of the missing value(s) is performed by using the values available at the sensors relating to the MS through association rule mining. Temporal data mining using data clustering is also employed to improve data estimation. This research derives solutions for both centralized and distributed wireless sensor networks where transmissions can be single hops or multiple hops, and sensors/servers can be static or mobile. It then conducts performance evaluations using NASA sensor data to compare its proposed technique with existing statistical approaches. |
Bibliography
| | Research issues in data stream association rule mining
Jiang, Nan; Le Gruenwald; ACM SIGMOD RECORD, (March 2006) 35 pp. 14-19
| |
|
Liu, Biao; Master's Thesis - School of Computer Science - University of OKlahoma, (May 2006)
| |
| Estimating missing data in related sensor data streams
Halatchev, Mihail; Le Gruenwald; International conference on management of data, (January 2005) pp. 83-94
| |
|
|
