| Summary: | 碩士 === 元智大學 === 通訊工程學系 === 106 === In the navigation domain, Indoor Positioning have become the a principal focus to researchers and technology vendors. Promising approaches like the use RFID, UWB, Wi-Fi and BLE have been proposed to better enhance the technology. How- ever, a conventional Indoor Positioning System(IPS) is still a work in progress. For the purpose of this Thesis, an evaluation of of different algorithms is done on the Bluetooth Low Energy(BLE) approach using Received Signal Strength(RSS) as the root parameter. Moreover, iBeacon and Raspberry Pi 3 are two central tools used for the RSS extraction, which make up the corpus. To suffice the purpose of the experiment, statistical(Skew-normal & Gaussian Mixture model) and Machine Learning(Multilayer Perceptron) algorithms are reviewed then compared in terms error rate. At the end of the experiment and evaluation, substantial accuracy rate were achieved
|
|---|
