Performance Evaluation of WLAN Access Points Selection Metrics for Fingerprinting based Localization

abstract Reliability and cost are essential elements to consider in all engineering network designs. In RF-localization systems, reliability can be defined as seamless coverage with precise and accurate localization. Cost-efficiency aims to reduce the infrastructure while simultaneously maintaining high accuracy. Both the cost and reliability can be associated with Access Points (APs) deployment. Therefore, it is paramount to study how to optimize the APs placement in RF-localization systems. To select the optimal AP configuration, different selection metrics are proposed. This paper investigates different AP placement and optimization strategies for WLAN fingerprinting indoor localization systems. Performance of selection metrics are evaluated experimentally in a realistic indoor environment. A fingerprinting database is developed by a grid spacing of 2m using 7 WLAN APs, by collecting Received Signal Strength (RSS) values using an Android smartphone app. The experimental results show, the AP configuration obtained by the metric combining the fingerprint difference metric with the geometric dilution of precision, results in high localization accuracy.