Abstract: Wireless indoor positioning has been widely adopted in extensive practice for its highly accurate and reliable characteristics. However, in order to achieve good positioning accuracy, positioning algorithms must be designed to be compatible with wireless positioning facilities. With the application of 802.11 ax protocol, the available bandwidth of wireless network transmission system will increase from 80 MHz to 160 MHz compared to the previous system. in addition, more access points are deployed within the indoor area, resulting in serious impact on high-frequency signal attenuation caused by interference and wall penetration for high-precision indoor positioning. In addition, the added 5.8GHz transmission signal lacks a common effective data set to support positioning functions, posing a great challenge for researchers of FTTR-based scenarios for positioning.
For indoor wireless communication network systems, this study proposes a ray-tracing model-based 5.8 GHz data set assembly method. The method includes simulating access points to generate reference signals, performing endpoint channel estimation, and generating frequency response images. The frequency response matrix is generated within the available bandwidth of FTTR with the help of existing WiFi positioning datasets. To achieve high-precision indoor localization, this study further proposes a deep neural network (DNN) computation method based on parallel path principal component analysis (PCA) preprocessing. The method includes the preprocessing step of parallel path PCA, the training process of the DNN network, and the user location calculation. The classification matrix is generated using principal component analysis (PCA) of parallel paths by using a fully connected neural network for training to improve the localization accuracy. Experimental results show that the proposed localization algorithm achieves a localization accuracy of less than 1 meter, which is not only more accurate than the traditional location estimation algorithm, but also meets the demand for fine-grained localization in practical applications.
Keywords: Indoor positioning, Wi-Fi, FTTR, PCA, DNN networks.
|
DOI:
10.17148/IJARCCE.2024.131101
