Research On Data Fusion Algorithm Based On Multi Physical Parameters Of Body Area Network

With the rapid development of telemedicine monitoring,body area network comes into being.It is a wireless sensor network applied to the monitoring of human body's multi-sign parameters.Since its inception,wearable domain networks have been used to achieve continuous and real-time monitoring and feedback at low cost,bringing many conveniences to people's daily life.Body area network is widely used in big data medical treatment.Its core technology is the fusion of multiple sign parameters.By data fusion algorithm,we can deduce the current physical state of the human body so that medical staff can intervene and treat in time.However,there is no uniform standard for body area network data fusion technology,and how many physical parameters can fully reflect the status of physical signs has become a research hotspot.In view of this,in order to better fuse the sign parameters and improve the recognition rate of human sign states,the following studies have been done on data processing and data fusion of body area network respectively:(1)The mechanism,reading and anomaly detection of multi-sign parameters in body area network.The change of sign state is reflected in the fluctuation of sign parameters,so the mechanism of sign parameters needs to be analyzed.In order to better test the performance of the data fusion model,this paper uses the physical parameters collected by the domain network set up by the U.S.intensive care unit,reads them through software programming,completes the visualization of the data,more intuitively clarifies the temporal variation and correlation between the physical parameters,and paves the way for anomaly detection.Because this dataset has a large number of outliers caused by sensor faults,it needs to be removed.Therefore,this paper presents an outlier detection algorithm for body domain network,which has achieved good results in this study,taking into account the mechanism and characteristics of multi-sign parameters.(2)Data fusion model based on improved fuzzy neural network.In order to find a suitable data fusion model for the body area network,this paper describes and analyses several common data fusion methods,points out their mismatch and inappropriateness with the body area network,and finally establishes the fuzzy neural network as the data fusion algorithm in this paper.In order to solve the problem that the fuzzy neural network is easy to fall into local minimum and over-fitting,this paper improves the particle swarm optimization algorithm,applies the algorithm to the optimization of parameters and weights of the fuzzy neural network,and applies the model that obtains the optimal combination of parameters to data fusion.By simulating the real data of the body area network,we can see that the data fusion model used in this paper has higher fusion accuracy.In the application of multi-sign parameters in the domain network,the anomaly detection algorithm and data fusion model proposed in this paper can avoid false positives caused by errors in a single sensor,improve the reliability of the body sign information fusion in the domain network,realize the perception and accurate judgment of the condition of the body sign,and then make a quick response.Therefore,this study has some reference significance and application value.