Research On Wireless Off-body Channel Propagation Characterisrics And Sparse Modeling

In the 5th generation and beyond the 5th generation mobile communication networks,a large number of small-size devices with low complexity,low cost and low power consumption will be widely connected.The portable devices worn on human body(animals,robots and other moving objects)are one of the most important components.The successful application of such wearable devices relyes on the research on the propagation characteristics of off-body channels which plays a key role in connecting the wireless body area network and cellular network,WiFi and other local area networks.The off-body channel is defined as the transmission medium between the wearable device and a remote access point(AP).As the off-body channel is characterized by complex and versatile scenarios,obvious antenna-body effects,being vulnerable to human gestures and motions,its propagation characteristics are significantly different from traditional short range wireless channels.It's of great theoretical significance and practical value to comprehensively understand the theory of off-body propagation so as to construct an efficient and robust off-body access link.This thesis concentrates on the scientific problem of the analysis and modeling of the off-body propagation characteristics.Large amounts of measurement champions are carried out in the typical channel types of single link,diversity reception,changing body worn positions,changing the heights of access point and multiple human gestures.Then,systematic analysis on the measured big channel datasets are conducted based on traditional large/small scale propagation analysis methds and compressive sensing based sparse channel analysis methods.Some large/small scale off-body channel models with high prediction accuracy and versatility are established;a highly robust circularly polarized spatial diversity off-body scheme is proposed and validated.Finally,the significant advantages of sparse channel modeling in single and multiple channel modeling are clarified.The main work and innovation points are as follows:(1)A two-factor integragted path loss model with variable body worn locations and variable AP height is established to provide a solution of accurate path loss predition for off-body channel under complex hospital scenarios.The power loss exponents are modeled as a quadratic function of AP height factor,and the body-worn position loss is considered as a look-up table for covinent entension and easy parameter update.Comparing to traditional large-scale models,the proposed two-factor path loss model is equipped with wider application range,stronger scenario adaptability and deeper insights.Numeric simulation shows that it predicts path loss more accurately than the logarithmic distance height gain model.(2)A highly reliable circular polarization diversity access scheme is proposed,and a diversity signal model considering the combined gain and polarization mismatch loss is established to provide a robust access solution for wearable networks.By comparing the measured propagation characteristics of traditional Patch and PIFA linear polarization diversity scheme and the proposed circular polarization scheme as well as the theoretical derivation of the diversity signal model,it is found that the proposed circular polarization scheme has significant advantages and can counter the serious receiver and transmitter polarization mismatch loss in wearable communication.Monte carlo simulation further verifies the advantages of the proposed scheme and found that the average received signal strength increased by 2.1dB.(3)The pulse response model of isolated channel based on single measurement vector compressive sensing(smv-cs)is established,and the compressive sensing framework was innovatively introduced into the channel modeling field to solve the problem that the description of TDL model is not accurate enough and the statistical modeling of SV cluster model is too complex and inflexible.The three balance principle of modeling complexity,precision and sparsity is established.According to this principle,appropriate dictionary of sparse analysis,recovery algorithm and model accuracy selection are carried out.Based on CS rich algorithm library and wavelet dictionary library,the measured wearable channel is modeled for channel impulse response.Thus,a sparse statistical analysis and modeling framework based on smv-cs method is constructed,and a mathematical model of the bridge relationship between channel sparse parameters and propagation of first-order and second-order statistics is established.According to the exponential attenuation characteristics of the channel sparse coefficient vector,three methods of sparse channel modeling for SMV-CS are proposed.Through a large number of Monte Carlo simulations of the measured multi-scene off-body channels and the simulation channel recommended by IEEE802.15.6,it is found that the proposed three models have obvious improvement over the statistical TDL model in complexity and accuracy.(4)An off-body synchronous sparse channel model based on multi-measurement vector compression sensing technology(MMV-CS)was established to extend the research results of the previous section from single channel to multi-channel synchronous measurement application,so as to solve the modeling problem of high-capacity and high-efficiency off-body networks.In order to meet the urgent needs of high-quality pictures and video acquisition,based on the 3.65 GHz continuous bandwidth allocated by WBAN standard in UWB,the design of large scale multiple channels or high bandwidth channel supporting high burst wearable communication requirements cannot be ignored.However,such high-dimensional off-body channels depend on the characteristics of synchronous channels and the construction of synchronous channel modeling method.MMV-CS provides a new solution.In order to use the CS method to effectively extract the co-supported dictionary set of synchronous channel,the algorithms based on wavelet soft threshold denoising,synchronous channel enhancement measures and adaptive dictionary selection are designed.A synchronization enhancement matching tracking algorithm(TA-SOMP)is implemented to carry out synchronous sparse analysis using wavelet dictionary,and a synchronization sparse channel modeling method is constructed.Compared with traditional modeling methods,it has many advantages,such as low complexity,support for real-time modeling,wide application of scenarios,and great potential for improving accuracy.The propagation characteristics of in off-body channels and sparse modeling methods studied in this paper all have corresponding models and sparse analysis modeling simulators,whose source codes are shared on gitHub.It is expected to provide useful theoretical basis and engineering reference for the coverage planning,link simulation,algorithm design,performance analysis and modeling by using advanced sparse learning tools in the field of wearable network interconnection and in off-body access in the future mobile communications.