Indoor Fingerprint Localization Algorithm Based on Variable Weight Using Channel State Information

Abstract

In order to solve the problems of indoor fingerprint localization algorithm based on static weight, such as low positioning accuracy and poor environmental adaptability, a variable weight indoor based on channel state information with Euclidean distance as weight reference is proposed. In the preprocessing stage, the collected CSI amplitude values are first subjected to Butterworth low-pass filtering denoising processing, and then the values of each sampling point are averaged, and the reference point location fingerprint database is established by combining the known coordinates. The weight index α is introduced in the online positioning stage, and the nearest neighbor is found by the KNN algorithm with the CSI eigenvalue reference. Then the weight index β is introduced, and the Euclidean distance is used as the weight reference, and the nearest neighbor is weighted. Get the coordinates of the target position. The experimental results show that the indoor fingerprint localization algorithm proposed by this scheme has higher positioning accuracy and less fluctuation of positioning error than the traditional KNN-based indoor fingerprint localization algorithm.