RSS based indoor positioning approach in indoor environments: Implementation of fingerprinting method and its performance analysis

Wi-Fi technology, which provides wireless communication, has become a tool that is widely used today, updated and renewed every year and benefiting human life in various fields. Although the purpose of its emergence is wireless communication, as a result of the studies carried out by the researchers, Wi-Fi signals are also used for positioning as a side task. The fact that wireless location determination is a very popular and up-to-date research subject is due to the fact that this technology is quite common throughout the world and that the infrastructure is ready in almost all public buildings, shopping malls, hospitals and libraries. Signal fingerprints are the signal vectors created by many wireless Wi-Fi access points in these buildings, and as the name suggests, these fingerprints are specific only to the location they belong. In this thesis, the received signal strength (RSS) information and fingerprint positioning method were applied in real conditions, and their position accuracy was analyzed in terms of various criteria. For this purpose, a data collection software, 3 proposed data types and an analysis software were developed. With the measurement mechanism used for the calibration phase, which is the most time-consuming phase of the fingerprint method, and the calibration points created in the form of routes, it was ensured that the fingerprint database was created with high accuracy and quickly. In our studies, it was seen that the positioning accuracy of the nearest neighbor method is highly dependent on the calibration point density. It is understood that weights should be taken into account in the calculation of candidate point locations using KNN algorithms, but it would be more appropriate to use a dynamic value instead of a fixed value in the selection of the number of neighbors for different types of indoor spaces. It is stated that if there are data gaps in the radio map depending on the frequency of the point, the algorithm can match the position of the candidate point with the calibration point that is physically far from that point, therefore filling these gaps with interpolation methods will benefit position accuracy. Software solutions that can be used in RSS based fingerprint indoor positioning applications accelerate the process considerably and enable various analyzes to be performed with desired parameters. The analysis and data collection software created with this thesis are projects whose development processes are ongoing and have become two programs that meet the basic needs when an indoor positioning application is needed. It is thought that these programs will help in solving many problems that need to be investigated and analyzed under real environment conditions such as the problem specified in Section 4.1. With the Wi-Fi 6 technology that emerged at the end of this thesis, it is undoubted that new horizons will be opened in indoor positioning studies with frequencies of 6GHz and above. In addition, the Covid-19 pandemic that occurred at the end of the study showed that in such cases, location information and determination of contact with people at risk play a very important role in determining the conditions such as exposure to the virus and risk status.