Indoor Fingerprint Positioning Based on WiFi Channel State Information

Since WiFi is the most widely utilized form of communication in daily life and has been installed in almost all public infrastructures, high-precision indoor positioning based on WiFi offers a lot of growth potential. The received signal strength (RSS), which is frequently employed in WiFi-based indoor positioning, is substantially impacted by the multipath effect in a complex environment, making RSS only viable for coarse-grained indoor positioning due to the high complexity of the interior environment. Therefore, in this paper, we present a fingerprint positioning method based on modified Adaptive Boosting (Adaboost) algorithm in order to accomplish high-precision positioning by designing a positioning system using WiFi channel state information (CSI) rather than RSS. In the offline phase, the CSI features for the generation of the fingerprint database are selected using the Chi-Square test and a modified Adaboost algorithm is applied to create a nonlinear mapping between CSI features and position. In the online phase, an improved probability-weighted algorithm is used to accurately estimate the position of the point to be measured. This paper evaluates and compares the placement performance of this system in the experiment to other systems. The experimental findings demonstrate a significant improvement in positioning accuracy with the positioning system suggested in this paper.