In view of time-varying

non-linear and strong correlation characteristics of multi physiological parameters in wireless body area sensor networks（WBANs）

in order to effectively improve the accuracy of medical monitoring system in early warning of patient’s physiological changes

a data fusion algorithm for body area networks is proposed.The fuzzy neural network（FNN） is used to establish the decision model for data fusion

and the improved particle swarm optimization（PSO） algorithm is applied for optimizing of parameters of the fuzzy system.Through the network training and testing of the real data of ICU in the PhysioBank database

the results show that the data fusion model can effectively process complex physiological parameters

and achieve high detection rate and low false alarm rate.