Indoor Localization Algorithm Based on Artificial Neural Network

Abstract

Indoor environment is complex and changeable, and the application scenario of indoor positioning(IP) requires higher accuracy, while the technology of outdoor positioning cannot meet people's requirements for IP, so IP technology is born. Since artificial neural network(ANN) are highly adaptive, fault-tolerant and suitable for complex indoor environment analysis, this paper introduces neural networks to solve the indoor localization problem. After the simulation study of ANN localization algorithm(LA) in this paper, the localization error accumulation probability and root mean square error of localization are compared between ANN localization model and other classical LAs, and it is found that compared with several LAs such as KNN, MLP and SVR, ANN LA has better performance and higher localization accuracy under low signal-to-noise ratio.