碩士 / 國立高雄應用科技大學 / 資訊管理研究所碩士班 / 101 / The geometric dilution of precision (GDOP) concept was originally used as a criterion for selecting the optimal geometric configuration of satellites in global positioning system (GPS). If the measurement variances are not equal in each other, we can choose the weighted GDOP (WGDOP) instead of GDOP, which could be the most appropriate selection criterion of location measurement units. The conventional matrix inversion method for WGDOP calculation has a large amount of operation and requires long time for computing, which would be a burden for real time application.
In this research, the original six kinds of input-output mapping based on resilient back-propagation (Rprop) are extended to WGDOP based on support vector regression (SVR). From simulation results, the proposed architectures for WGDOP approximation based on SVR always yield superior estimation accuracy. This research use a set of four base stations (BS) selected from seven BS to estimate mobile station (MS) location in cellular communications system. The mobile positioning of a wireless network plays a key role in providing location-based services. Different applications of location-based services have been well developed, including E-911 subscriber safety services, intelligent transportation system (ITS) and medical treatment system. To further reduce the complexity, our approach is to first select the serving BS and combines it with three other measurements to estimate MS location. As such, the number of subsets is reduced greatly without compromising the location estimation accuracy. With the minimum WGDOP algorithm, MS location can be estimated by the linear lines of position algorithm (LOP) and distance-weighted method.
| Identifer | oai:union.ndltd.org:TW/101KUAS0396020 |
| Date | January 2013 |
| Creators | Chao-Yi Wu, 吳昭儀 |
| Contributors | Pei-Yi Hao, 郝沛毅 |
| Source Sets | National Digital Library of Theses and Dissertations in Taiwan |
| Language | zh-TW |
| Detected Language | English |
| Type | 學位論文 ; thesis |
| Format | 62 |
Page generated in 0.0018 seconds