| Summary: | 碩士 === 國立臺北教育大學 === 資訊科學系碩士班 === 99 === In recent years, with the rapid development of wireless sensor network, the applications of location-based services have been increased. Therefore, localization in wireless sensor networks has become a popular research issue. Many localization algorithms for WSNs were proposed. Among them, approach based on received signal strength (RSS) is simple and economic since it needs no extra hardware. However the received signal strength is greatly influenced by the surrounding environment, and hence leads to inaccurate distance estimation. This thesis proposed a collaborative localization method to reduce the errors while using received signal strength for location estimation. In this approach, a bind node will periodically broadcast packets for collecting the RSSI values between it and its neighboring nodes. Moreover, by using the shortest path algorithm, the distance between the blind node and the reference node can be better estimated. Thus, with at least three references nodes, the location of a blind node can be computed by using Minimum Mean Square Error algorithm. In order to increase the accuracy of localization, our thesis designs a filter that the unreliable RSSI values are discarded, and the distance by using the shortest path algorithm is modified by a ratio constant. Our experiments are carried out with TI CC2430/2431 chips. The experimental results show that, the estimated position is very close to the actual position of the blind node. That is, our proposed mechanism can effectively reduce the errors in location estimation caused by using weak RSSI signals.
|
|---|
