Human Body Detection Of Wireless Sensor Network Based On RSSI

Wireless sensor network (WSN) is a special peer-to-peer network, and its huge number of are widely distributed. The research and development of WSN based on ZigBee technology has received more and more attention at present. WSN need low transmission delay and low power consumption but not a high transmission bandwidth which users can have a longer battery life and more device array. ZigBee has many advantages such as high communication efficiency, low complexity and low power consumption, low rate, low cost and so on. These advantages make ZigBee and WSN combine perfectly.This paper does human body detection based on ZigBee system. Firstly, the hardware and software of ZigBee system was designed. The radio frequency chip CC2500was used for data packets remote transmission, and transferred packets to the host computer through the serial port communication, then calculated Packet Reception Ratio(PRR) to evaluate network link quality. However, with network link quality is dynamic along with propagation environment change, PRR based on a large number of data packets cannot reflect the real time link quality accurately. So we raised a real-time RSSI link quality evaluation method based on exponential smoothing. PRR is computed in real time by using the window average time and the value of smoothing constant is decided by statistical character of RSSI of the proposed method. And then the method was verified by a large number of experiments.Secondly, we study indoor wireless electromagnetic wave propagation, floor reflected power curve characteristics is analysized using the knowledge of emission coefficient and transmission coefficient,"3σ" law is used to set the loss of indoor electromagnetic wave propagation when encounters obstacles as the Gaussian random distribution and hypothesis test is made to the theory of variance. People’s influence on electromagnetic waves is also discussed by the same theory. The transient behavior of RSSI in no human body environment was analyzed by Auto-Correlation Function(ACF).The Human body detection based oh RSSI just need the RSSI volatility and did not need external sensor. The proposed method filtered the fluctuation caused by environment by using Intermittent Karman in data packets loss situation. We analyzed the human body influence to network link quality by Signal Noise Ratio(SNR)which computed through Karman filter, then the human body detection procedure is introduced, and a large number of detection experiments verified the method practicability and extensibility. Finally, Elliptic attenuation model was used for further detection.In short, human body detection based on RSSI is discussed in this paper, some preliminary results are obtained and the research of the volatility of RSSI is worth for further attention.