以無線網路為基礎之定位服務是目前相當熱門之研究議題;它使用目前最普及之802.11標準,不但可以利用現有之無線網路環境以有效地降低建置定位服務之成本,更可以達到定位系統之廣泛應用。本研究分析目前可行之無線區域網路定位服務相關技術,並針對室內環境定位服務問題進行探討。 為了達到能在室內空間定位並提供適當之資訊服務,本研究提出三角定位及特徵定位方法,利用自行設計之使用者操作介面,針對定位環境之訓練點進行訊號強度實測,建立所有訓練點之訊號強度特徵資料庫。藉由行動裝置即時收集每台AP之訊號強度值(RSSI),主要以特徵定位法進行定位運算並與所建立之資料庫進行比對,進而判斷出該使用者之位置,應用於自動導覽系統。實驗結果顯示以SimNet相似度比對方法相較於Inner Product與Nearest Neighbor方法的準確度為最高。另外,由於室內環境干擾的因素,三角定位法將會產生極大的誤差範圍。
The positioning service that based on wireless network is a quite hot research topic at present; it uses 802.11 standards that popularized the most recently. It not only utilize the existing wireless network environment to reduce the cost that the construction oriented the service effectively, but also reach the wide application of the positioning system even more. The research analyzes feasible WLAN positioning techniques, and carries on the discussion to the indoor positioning problems. In order to find out the position and offer proper location-based information, we proposed both triangulation and fingerprinting methods. Both methods measured the signal strength of every APs' RSSI for further calculation. However, the fingerprinting methods needed some pre-trained data points for positioning. The experimental results had shown that the fingerprinting method which utilized SimNet function yielded highest accuracy comparing with Inner product function and Nearest Neighbor function. Moreover, the triangulation method yielded worst results mainly due to the complicated indoor environment.