Improving Wi-Fi Indoor-Positioning Accuracy by Using AP Selection and Adaptive Pattern Matching

[[abstract]]This study proposes a method based on access point (AP) selection and adaptive pattern-matching for Wi-Fi indoor positioning (ASAPM). In the proposed ASAPM, a box plot algorithm is used to remove received signal strength (RSS) outliers in samples received from APs in order to smooth the RSS. Subsequently, we analyzed the RSS variations for selecting the top-N APs with the least interference. Moreover, we analyzed the history of the positioning results to estimate the direction and distance of users in subsequent positions in order to reduce the pattern-matching time and computational overhead of the positioning system. The simulation results revealed that the average positioning error, average maximum positioning error, and average pattern-matching times of ASAPM were 36%, 51%, and 57% lower than the three compared strategies, respectively. These findings show that ASAPM could reduce the computational overhead; moreover, it is suitable as an indoor-positioning service for mobile devices