Abstract: Aiming at the problem of equipment heterogeneity and the change of Bluetooth beacon nodes in indoor location based on Bluetooth fingerprint, we proposed an indoor location algorithm for heterogeneous devices and environmental changes. Firstly, we used Procrustes analysis method to standardize the received signal strength, and used kernel extreme learning machine (KELM) to model the standardized fingerprint database to reduce the signal strength differences caused by the differences of users’ mobile terminals. Secondly, when the access point (AP) signal changed, the access point signal was recalibrated by Gaussian process regression (GPR), and the fingerprint database was updated to eliminate the positioning error caused by the weak signal, position movement or environmental change of the access point. Test analysis results show that the algorithm can effectively overcome the impact of heterogeneous equipment, and better adapt to the environmental changes.

Key words: indoor location, Procrustes analysis method, heterogeneous equipment, environmental change, Gaussian process regression (GPR)