There is a limitation of IPS (Indoor Positioning System) because of unstable GPS (Global Positioning System) signal in indoor space. The signal is unstable because of the barrier (walls) between the user and the satellites. This study develops EIPS (E...
There is a limitation of IPS (Indoor Positioning System) because of unstable GPS (Global Positioning System) signal in indoor space. The signal is unstable because of the barrier (walls) between the user and the satellites. This study develops EIPS (Enhanced Indoor Positioning System) to overcome that limit. The EIPS combined Wi-Fi fingerprinting, Magnetic Fingerprinting and Image Processing technology to locate and navigate user in indoor space.
The EIPS consists of four modules: WFPM (Wi-Fi Fingerprinting Positioning Module), MFPM (Magnetic Fingerprinting Positioning Module), FPM (Final Positioning Module) and PPM (Picture Positioning Module). The WFPM collects the Wi-Fi fingerprinting data. The WFPM processes the collected data with MANN to obtain the user predicted location-1. The MFPM collects the magnetic fingerprinting data and processes the collected data with a DTW and MMCL algorithm. The FPM receives predicted location-1and predicted location-2 of the user from WFPM and MFPM. The FPM calculates the final predicted location of the user. The final predicted location is stored to the central database. The PPM makes the user take the picture of landmarks around him. The picture is processed by the ESVM to help the map to be generated. The user also can search for their location or destination more accurately by using the PPM. The result of the PPM is compared with the final predicted location and stored in the Central Database.
The EIPS is simulated using MATLAB. The simulation data is collected using iPhone 6 from the 3rd floor of Engineering Building, Catholic Kwandong University. The RPs are placed all over the corridor one meter apart from each other. The magnetic fingerprinting is collected by walking around the corridor.
In the first simulation, the Wi-Fi fingerprinting obtained about 89% accuracy, the magnetic fingerprinting obtained about 93% accuracy and the EIPS obtained about 94% accuracy. The EIPS provides more accuracy than Wi-Fifingerprinting and magnetic fingerprinting. In the second simulation, the Wi-Fi fingerprinting obtained 3.6 meters average error, the magnetic fingerprinting obtained 3.1 meters average error and the EIPS obtained 2.9 meters average error.Hence, the EIPS has less average error than Wi-Fi fingerprinting and magnetic fingerprinting. This study also showed the effectiveness of Gaussian Filter which is used in the PPM and showed that the ESVM provided more accurate result and less error rate than SVM.