Relay Node Placement in Wireless Sensor Network

碩士 === 國立清華大學 === 資訊系統與應用研究所 === 98 === The wireless sensor network (WSN), is by many low costs and low power sensor node (SN). The first use of WSN is American Military for data collection. SN can perform sensing and transmitting sensed data to others. However, the prime task of SN is sensing, not...

Full description

Bibliographic Details
Main Authors: Lee, Yi-Ting, 李宜亭
Other Authors: Shih, Wei-Kuan
Format: Others
Language:en_US
Published: 2010
Online Access:http://ndltd.ncl.edu.tw/handle/15922722334963111213
Description
Summary:碩士 === 國立清華大學 === 資訊系統與應用研究所 === 98 === The wireless sensor network (WSN), is by many low costs and low power sensor node (SN). The first use of WSN is American Military for data collection. SN can perform sensing and transmitting sensed data to others. However, the prime task of SN is sensing, not transmitting. It is not suitable for SNs to transmit their data with long distance. A relay node (RN) is a node which is costly, but more powerful than SN. The main function of RN is to communicate with other nodes. RN sends the sensed data to the base-station via multi-hop. How to place the minimize size of RN in wireless sensor network is our major goal. We assume the communication range between SNs is r and between RNs is R, and 0<r<=R. Notice that SN can just send the sensed data to the RN. Then, RN transmits those data by multi-hop. Our goal make the sensor network connected with the minimal size of RN. This problem can be divided into two sub problems. The first sub problem can be solved by geometric disk cover algorithm, and the second sub problem can be solved by STP-MSP algorithm. It has been shown that the Minimum Geometric Disk Cover problem is NP-complete. Thought there is a PTAS for the problem, it really cost time when implement it. Therefore we use some trick to make it faster. We also translate this problem to set cover problem for saving time. The STP-MSP is NP-hard and there is a 3-approximation algorithm. This thesis presents two frameworks with three principles. Given different parameters, we can reduce size of RN to 96% in average. In some specific parameters can reduced to 90%. For some special case we can reduce 50%.