Bluetooth Indoor Micro-positioning System based on Gaussian Filtering and Elman Neural Network

Abstract

Due to the complex and diverse indoor environment, the traditional positioning algorithm used indoors will greatly reduce the positioning accuracy. Aiming at the ever-increasing demands of accuracy and stability in indoor positioning, an indoor micro-positioning system based on Bluetooth technology is designed. The low-power Bluetooth wireless data transmission module RL-CC2541-S3 is used as the signal acquisition unit, and the STM32F103ZET6 high-performance microprocessor is used as the core to complete the data processing and positioning calculation. An indoor positioning algorithm based on Gaussian filtering and Elman neural network is designed to improve the positioning accuracy of the system. Firstly, the Pauta criterion is used to eliminate the outliers of the sampled data to ensure the validity of the data. Then Gaussian filtering is used to remove the interference of Gaussian noise and the least squares method is used for curve fitting to further improve the accuracy of ranging. Finally, the nonlinear approximation of Elman neural network is used to achieve the target location. The experimental results show that the absolute error of indoor positioning is close to 0.3m, which is better than the traditional Chan algorithm, LS algorithm and Taylor series positioning algorithm, which can meet the needs of general indoor micro-positioning.