Adapted WLAN Fingerprint Indoor Positioning System (IPS) Based on User Orientations

Abstract

Location determination techniques for IPS are divided into proximity, triangulation, and fingerprint techniques. WLAN-Fingerprint has become one of the best alternative methods because it is already widely available inside buildings and provides a high level of accuracy. However, the WLAN signal is highly influenced by the surrounding environment, such as walls, ceilings, and people. People who hold mobile devices (MDs) significantly affect IPS performance. This paper therefore proposes a new method to overcome the effect of user orientation on the accuracy of IPS by introducing an adaptation of the signal strength values based on user orientation. The average position error when using the initial radio map (RM) is found to be 8.84 m. After implementing the RM adaptation, the average error decreased to 3.5 m. This proposed method can therefore improve accuracy by 60.4% and decrease the required RM database memory with more additions of simple computations.