Research On CSI-based Indoor Positioning System Adapted To Environmental Changes

DFL systems have aroused extensive attention because it is more convenient than DEL systems,and fingerprint-based localization method is usually used in DFL systems.Although fine-grained information can be provided by the CSI,changes in the environment over time can cause the CSI become different.Therefore,the real-time CSI data can't match with the data in the fingerprint map established beforehand very well,which can lead to the inaccuracy of the positioning result.This paper presents a DFL system,which adopts TL method to update the fingerprint map and employs the Light GBM algorithm to train the fingerprint map.The main work of this paper is as follows:(1)Studying a positioning classification algorithm based on Light GBM as the classification method.Light GBM algorithm uses the leaf-wise growth strategy to improve the localization performance.Compared to other EL algorithms,like GBDT and XGBoost,it can increase the accuracy of positioning.(2)Researching on applying the TL method to resolve the problem caused by the instability of the CSI data owing to the environmental changes.Specially,for the sake of reducing the overhead caused by re-collection of the CSI data,we apply JDA algorithm to find the correlation of the CSI data between the target and source domain,thereby updating the fingerprint map established beforehand to alleviate the effect brought by the change of CSI.(3)Testing the proposed location method,and the performance of the proposed method is evaluated according to the characteristics that CSI fingerprint database would change with time and environments.In order to verify the effectiveness and robustness of the proposed method in different environments,two different positioning scenarios are selected for verification.Particularly,the robustness of the proposed approach was evaluated by using the CRLB as the standard of measurement.The experimental results show that the average distance error after applying the calibration method can be closer to the root of CRLB,which indicates that the system has strong robustness,and can effectively reduce the overhead caused by re-collecting CSI data during the update stage while maintaining the localization accuracy.Meanwhile,The results of experiments demonstrate that the positioning accuracy can be improved by using JDA to update the fingerprint map established before.