Indoor Positioning using Wifi Fingerprints

Abstract

Indoor Localization has emerged as an important component for a variety of indoor appli cations, such as wireless advertising, pedestrian navigation, and information retrieval. The need for very precise and affordable indoor positioning has attracted a lot of attention from the scientific and business worlds. In particular, identifying a mobile user’s floor level is essential for a variety of location-based applications in situations that include multi-story structures. Numerous indoor localization methods have been developed as a consequence, utilising open source datasets that include Wi-Fi fingerprints and Machine Learning and Deep Learning models. These techniques, which rely on measurements from Wireless Access Points (WAPs), em ploy the Received Signal Strength to guess where mobile users are located using an existing fingerprint database. To get the best results, the dataset is normalised using two equations, and each machine learning and deep learning model has its hyperparameters tuned. By addressing the demand for accurate and effective location-based services in complex in door environments, the employment of these localization techniques offers potential options for indoor positioning. Wi-Fi fingerprint data is used by machine learning and deep learning models to produce precise and dependable indoor localisation.