A Data Fusion Algorithm Based On Time Series Prediction Method For Wban

The development of wireless communication technology has completely changed our daily life,its technical applications involve automatic control,tracking and monitoring.With the development of wireless sensor network(WSN)technology,the Wireless Body Area Network has presented itself.By wearing and even embedding low-cost,energy-constrained,tiny sensor nodes,make up a special type of wireless sensor network,to collect physiological information,transmit the information to the central station through wireless communication,and thus realize real-time monitoring of various indicators of human body.WBAN devices provide real-time feedback,won't cause any discomfort to human body,provide users with more flexibility and mobility,can substitute for complicated cable medical equipment,and continuously monitor the important physiological signals and entertainment signals.As all other electronic systems,WBAN also needs proper power to ensure its normal operation.Currently,the WBAN is mostly battery-powered.But compared with other WSN system,the sensor network design and application environments of WBAN is special,the batteries are difficult to change in time,so using of all kinds of energy saving technology,to provide a very long network life is one of the key areas of the WBAN study.Generally speaking,WBAN's strategy is to use more energy-efficient electronics,reduce the amount of devices,and adopt more efficient algorithms.For a given battery capacity,increase the energy efficiency of the system.The energy consumption of WBAN can be divided into three parts: perception,wireless communication and data processing,among them the wire less communication is the most power-hungry.The information collected by the WBAN is typically high in time redundancy and space redundancy.By using data fusion technology,reducing the data redundancy and the amount of wireless traffic in WBAN can achieve the goal of saving energy.Based on the analysis above,this paper has done the following work.(1)This paper presents a data fusion algorithm based on wavelet analysis and least-squares support vector machine.By establishing a similar lightweight dat a prediction model in the sensing node and the aggregation node,forecast the next data in the same time,when the forecast results within the error range is not for data transmission,so as to reduces the amount of the redundant data transmission in the grid or reduce the energy consumption of wireless network.(2)This paper studies the methods of parametric optimization of the predictive model.According to the theoretical analysis,an improved GA-PSO algorithm has been proposed.The new algorithm improves the parameters of the original algorithm and combines the idea of genetic algorithm to overcome some shortcomings of the original algorithm.(3)This paper designs a WBAN system to monitor the basic physiological data of human.In this paper,the architecture of the system was designed,wireless communication protocol was analyzed and selected.We completed the hardware and software design and implementation of sensor nodes and gateway nodes.