Filtering and fingerprint matching methods for Wi-Fi radio map based indoor localization

摘要

Usually the indoor localization methods are using predictor-corrector methods for the precise location determination. The prediction is based on the analysis of the motion of the user, while the correction is based on a radio map-based fingerprinting. There are some application fields where the precision of the predictor-corrector algorithm is sacrificed for the simplicity of sensors and devices. The recent development of the smartwatches they became ideal devices for monitoring elderly people's health conditions together with their indoor location and activity. Since it is not a navigation like application the expected precision and sample rate of the location determination is lower, however the typical sensors configuration of a smartwatch is more limited compared to a mobile phone. Usually the sensors are required for the motion prediction methods are missing, so the common predictor-corrector methods, filters are not applicable for this type of use case. Different algorithms and filtering methods needs to be developed and applied for this application. This article summarizes our approach and results in developing fingerprint only indoor localization application.