基於信號強度值之定位方法下改善因人體影響所造成之誤差

Abstract

室內定位是近幾年熱門探討的重點，其中又以基於無線網路的定位技術，因為其成本及便利性，而得到眾多研究及應用的青睞。然而無線網路室內定位的精準度，深受無線電波信號強度值(RSSI)的影響，RSSI容易受到來自環境的干擾，如:反射、繞射干涉以及障礙物遮蔽，在對人進行定位時，需要讓其穿戴收發信號的裝置，此時無線電波的傳遞會受到人體遮蔽的影響，影響最後的定位結果。 本論文實際針對人體干擾進行了量測，當背對於節點時，收到的RSSI值會產生極大的衰減，直接使用三點定位法無法找到明確的定位結果。我們透過判斷人體正背面與各信號節點間的相對位置，來改善人體造成的影響，降低了定位的誤差，在只有一個節點受到人體遮蔽影響較嚴重時，可以降低平均約1.5公尺的定位誤差。

Indoor localization is a popular topic in recent years, especially WLAN-based localization because of its cost and convenience. Nevertheless, the accuracy of WLAN-based localization is mainly determined by the receive signal strength index (RSSI). RSSI is affected by many factors of surroundings: e.g. reflection of RF signals, NLOS to signal sources. In tracking localization of a person, the person must wear a device with RF transceivers. Due to human body shadowing of RF signals, the localization result is affected. In this thesis, measurement of the human body shadowing of RF signals is focused. When facing away from a transmitter, human body shadowing will introduce a decay of RSSI value which hampers the localization finding with trilateration. In this thesis, the localization error caused by the human body is reduced by first determining the relative position of the human body and transmitters. Through the proposed method in this thesis, the localization error can be reduced by up to 1.5m when only one transmitter was affected by human body shadowing.