Aiming at the problems of low localization precision and poor adaptive ability to environment of traditional indoor fingerprint localization algorithm

a new algorithm based on parallel chaos optimization online sequential extreme learning machine（ PCOS-ELM） is proposed. In offline phase

parallel chaos optimization algorithm（ PCOA） is used for searching the optimization parameters of hidden layer nodes in the extreme learning machine

which constructs an initial positioning model with high precision; In the online phase

use online sequential extreme learning machine（ OS-ELM） to make the new added location fingerprint data online dynamically adjust localization model so as to adapt indoor environment changes. The results show that PCOS-ELM localizaition algorithm has higher localization precision and better adaptive ability to environment.