Research On Indoor Location Of Single Base Station Based On WiFi Channel State Information

With the wide application of multi-antenna technology in WiFi systems,single base station positioning based on WiFi channel state information has become a research topic.AoA estimation and ranging are two key technologies for single base station positioning.The existing AoA estimation method has larger estimation error due to low signal-to-noise ratio in the case of non-line-of-sight or long-distance signal transmission,and the existing ranging method has a fixed path loss factor in propagation model has larger range error in non-line-of-sight environments.Therefore,this article takes single base station positioning based on WiFi channel state information as research object,focusing on AoA estimation under low SNR conditions and ranging under non-line-of-sight conditions.The specific research contents and contributions are as follows:(1)AoA estimation based on WiFi channel state information under low SNR conditions.In the case of non-line-of-sight or long-distance signal transmission,the accuracy of AoA spatial spectrum estimation and the accuracy of direct-path AoA recognition are reduced due to severe signal attenuation.To solve this problem,this paper firstly applies wavelet denoising to improve the spatial spectrum estimation accuracy,and then proposes a direct path AoA estimation method based on hierarchical clustering-Logistic regression,which improves the accuracy of identification of the direct path AoA at the same received signal strength.Experiments show that compared with the existing methods under the non-line-of-sight condition,the 1-sigma error of the AoA estimation based on the proposed method is reduced by 16.7%.(2)Ranging based on WiFi channel state information under non-line-of-sight conditions.The path loss factor is an important factor in determining the ranging accuracy.The diversity of obstacles at non-line-of-sight makes the estimation of the path loss factor difficult.To solve this problem,this paper proposes a method of distance estimation based on support vector regression for dynamic estimation of path loss factors.The influence of the diversity of obstacles on the loss factor and the relationship between multiple statistical characteristics of the received signal and the loss factor are studied.A regression model is established between the multiple statistical characteristics of the received signal and the path loss factor.Experiments show that compared with the existing methods under the non-line-of-sight condition,the 1-sigma error based on the proposed ranging method is reduced by 11.9%.(3)Based on the above research,a single base station positioning system prototype was designed and a CSI positioning data set was established based on this positioning prototype which has over 5 million samples,cover 60 LOS and various NLOS test points and 4 Group experimental parameters.Relying on the data set,the single base station positioning performance based on the proposed methods and the existing methods was evaluated and compared,and the influence of different experimental parameter configurations on the positioning performance was discussed.Under the non-line-of-sight condition,compared with the existing method,the single-base station positioning 1-sigma positioning error is reduced by 14.8%.Experimental verification demonstrates that the positioning performance of the single base station positioning system based on proposed methods is better than that of the existing methods.The proposed method is feasible.