RF based indoor positioning system

Although the Global Positioning System is a publicly recognized technology for positioning in the outdoor environment, it is ineffective in the indoor environment. For this reason, the search for effective solutions to indoor positioning still continues. Within the scope of this dissertation, it is aimed to develop radio frequency (RF) based, high-accuracy and low-cost indoor positioning approaches based on the fingerprint method. For this purpose, in addition to the existing indoor positioning datasets in the literature, a new dataset has been constituted and made available to researchers. In terms of selected performance criteria, the most suitable algorithm for three different indoor environments is determined by multi-criteria optimization technique. Hybrid fingerprints are defined using a combination of WiFi received signal strength and magnetic field measurements. It has been observed that the positioning accuracy is improved when the proposed hybrid fingerprint dataset is used with different classification algorithms. F-score weighted indoor positioning algorithm combining WiFi received signal strength and magnetic field measurements is proposed. It has been observed that the accuracy of the proposed algorithm is higher than that of the conventional algorithms. In addition, an improved indoor positioning approach has been proposed that uses WiFi signal strength and magnetic field fingerprints for more precise locating. With this approach, high accuracy position estimation can be done.