| Summary: | 碩士 === 國立中興大學 === 電機工程學系所 === 102 === The Received Signal Strength Index (RSSI) is widely used on indoor positioning systems due to the advantage of being able to be sampled without requiring additional equipment. However, the irregular fading phenomenon of RSSI is a critical issue for systems adopting this kind of method to face.
To address the above problem, we propose a pedestrian positioning algorithm based on a dynamic fingerprint database in this thesis. More specifically, we use wireless access points (APs) as anchors to provide wireless signals to users for positioning. In the beginning of the proposed algorithm, we construct a fingerprint database with known starting position information of users or the position information computed from the MDS-MAP algorithm. Then we continuously update the database with estimated positions computed and RSSI readings collected upon each step of users. In doing so, we can real-time update the dynamic database according to the vibration of wireless signals to resist the irregular RSSI fading. More importantly, we employ the DOI model to simulate the irregular fading phenomenon of RSSI to explore the positioning errors of variable DOI settings, variable numbers of unknown nodes, different distributions of unknown nodes, as well as two kinds of database constructing and positioning schemes. Based on the results of our experiments, our algorithm can efficiently reduce the impact of the irregular signal fading.
|
|---|
