Please use this identifier to cite or link to this item:
https://ptsldigital.ukm.my/jspui/handle/123456789/486831Full metadata record
| DC Field | Value | Language |
|---|---|---|
| dc.contributor.advisor | Rosdiadee Nordin, Assoc. Prof. Ir. Dr. | - |
| dc.contributor.author | Zahid Farid (P65764) | - |
| dc.date.accessioned | 2023-10-11T02:25:50Z | - |
| dc.date.available | 2023-10-11T02:25:50Z | - |
| dc.date.issued | 2017-08-02 | - |
| dc.identifier.other | ukmvital:96519 | - |
| dc.identifier.uri | https://ptsldigital.ukm.my/jspui/handle/123456789/486831 | - |
| dc.description | Accurate wireless localization or positioning is an essential requirement for today's location-based services. Global Positioning System (GPS) provides reliable outdoor positioning but falls short when operating inside buildings due to the attenuations introduced by walls and ceilings. In indoor environments, Wireless Fidelity (WiFi) based localization is an obvious alternative, but is sensitive to various indoor fading effects due to multiple reflections that hamper the localization accuracy. Similar to other short-range wireless technologies, such as Bluetooth and Radio Frequency Identification (RFID), the existing positioning solutions are still limited in their accuracy and robustness in locating mobile devices. The most popular technique used to calculate the position of nodes is WiFi fingerprinting, though relying on a single technology renders localization technique less effective, thus motivating the research towards hybrid positioning that combines two wireless technologies. Exploiting the capabilities of several technologies will result in better positioning accuracy. The main objective of this thesis is to improve the accuracy of indoor localization systems by utilizing a combination of two wireless technologies in a single hybrid system. The proposed hybrid indoor localization technique adopts fingerprinting approaches based on WiFi and Wireless Sensor Networks (WSNs) wireless technology. This model exploits machine-learning techniques for position calculation. Artificial Neural Networks (ANN) based on the gradient descent with momentum (GDM) back-propagation learning is considered. To improve the ANN performance, Genetic Algorithm (GA) is further applied to optimize the learning parameters. The proposed system was validated against the dataset from the UJIIndoorLoC as a benchmark. Two sets of experiments were conducted in a real indoor environment: one using the proposed hybrid system, and another one using the single wireless technology, that is WiFi or WSN. In both experiments, two dimension (2D) and three dimension (3D) measurements are taken into consideration. In the case of 3D, the measurements are collected at three different height values. This work also implements a novel spatial filtering approach to minimize the fading effects that are usually found in the collected data. The position accuracy based on the WiFi technology can achieve average localization error of 1.22 m (2D) and 1.90 m (3D), in terms of the Euclidean distance between the estimated and the actual positions. Further improvement was achieved by optimizing the ANN-based positioning calculations using GA, which yielded an accuracy of 1.18 m (2D). On the other hand, the proposed hybrid system improved the accuracy using the ANN-GDM algorithm, reducing the average localization error to 1.05 m in the 2D scenario. Applying GA-based optimization in this case did not incur any improvement to the accuracy. Compared to the performance of GA optimization, the non-optimized ANN-GDM performed better in terms of accuracy, precision, stability and computational time. The above results show that the proposed hybrid technique achieves higher accuracy in real-world indoor positioning applications.,Certification of Master's/Doctoral Thesis" is not available | - |
| dc.language.iso | eng | - |
| dc.publisher | UKM, Bangi | - |
| dc.relation | Faculty of Engineering and Built Environment / Fakulti Kejuruteraan dan Alam Bina | - |
| dc.rights | UKM | - |
| dc.subject | Indoor localization | - |
| dc.subject | Wireless Fidelity | - |
| dc.subject | Indoor positioning systems (Wireless localization) | - |
| dc.title | Higher accuracy hybrid wireless indoor localization using machine learning | - |
| dc.type | Theses | - |
| dc.format.pages | 159 | - |
| dc.identifier.callno | TK5103.48323.F347 2017 3 tesis | - |
| dc.identifier.barcode | 002720(2017) | - |
| Appears in Collections: | Faculty of Engineering and Built Environment / Fakulti Kejuruteraan dan Alam Bina | |
Files in This Item:
| File | Description | Size | Format | |
|---|---|---|---|---|
| ukmvital_96519+SOURCE1+SOURCE1.0.PDF Restricted Access | 620.12 kB | Adobe PDF |  View/Open |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.