| Summary: | 碩士 === 國立臺灣科技大學 === 電機工程系 === 103 === Applications and services for smart handheld devices have garnered considerable attention worldwide, as have applications for Indoor Location-based Services (Indoor LBS). Improving their indoor positioning accuracy is a significant challenge. In recent years, many indoor positioning studies have used radio frequency for positioning; however, a signal’s susceptibility to environmental interference can decrease positioning accuracy. Increasing positioning accuracy is a complex problem and remains to be solved.
Indoor positioning systems typically use only a signal transmitter (Wi-Fi AP) and receiver to acquire reference point data. The systems then use the K-Nearest Neighbor algorithm (KNN) or a Multilateration algorithm for positioning. Consequently, positioning accuracy is limited by both the positioning algorithm and the use of a single source for acquisition of reference point data. This study utilizes the novel Integrated Heterogeneous Networks Indoor Positioning System (IHNIPS) with low-energy Bluetooth Beacon to obtain positioning data. A fingerprint algorithm is then applied to revise the multilateration algorithm and define the fuzzy area to improve the accuracy of subspace selection. Last, novel threshold mechanisms are used to choose the transmitter for positioning and thereby improve current positioning accuracy.
Experimental results show that the average error distance is 1.29m for Wi-Fi AP and 1.33m for Beacon. And the average error distance is 1.21m in the proposed Integrated Heterogeneous Networks Indoor Positioning System. The IHNIPS outperforms both Wi-Fi AP and Beacon. Thus, positioning accuracy is improved.
|
|---|
