Nonlinear Dynamics Characteristics Of Natural Wind Field And Its Application In Odor Source Research

The leakage of the toxic gas is a serious threat to human life, even result in socialpanic. Therefore, carrying out the basic research of the odor source localization is ofimportance in promoting the harmony society. In the recent several decades, themobile robot equipped with wind speed/direction sensor is widely utilized in theresearch of the odor source localization.Wind is the primary factor in spreading the odor. Due to the complex nonlineardynamics of the near-surface wind field, it poses great challenge in analyzing thewind field signal. To deal with the motivating problem, we employ the nonlinearanalysis and complex network to study the change of natural wind field and theoptimization in odor source localization. The contribution of this dissertation ispresented as follows:Firstly, an integrated technique which combines a direct identification method ofchaos, i.e., the saturated correlation dimension algorithm with surrogate data method(an indirect identification method) is used to analyze the nonlinear characteristics ofthe near-surface wind time series，the result indicates that the near-surface wind speedand wind direction signals have chaotic characteristics.Secondly, the time sequence signals of the wind speed and direction in thevertical and level wind field are analyzed by the utilization of Lemple-Ziv complexity,approximate entropy, the degree distribution exponent of network and so on. It isconcluded that the complexity of the time sequence signal of the wind speed is overthe wind direction, the complexity of the wind speed signal in the vertical field is overthe level field, and the complexity of the wind direction signal in the vertical field iseven with the one in the level field.Thirdly, the structural dynamic characteristics of the time sequence signal of thewind speed and wind direction are explored by the utilization of the phase spacenetwork motif frequency. It is concluded that the heterogeneity of the phase spacenetwork motif frequency of indoor wind speed signal is over the outdoor wind speedsignal, while the phase space network motif frequency of the indoor wind directionsignal is even with the outdoor wind direction signal. At last, we utilize the wind speed/direction sensor to collect the data of the windspeed and direction in50Hz, and then sample the collected data in7groups of timesequence by1Hz,2Hz,4.2Hz,5Hz,8.3Hz,12.5Hz and16.7Hz respectively. Based onthe7groups of time sequence, recurrence plot and recurrence quantification analysisare carried out to obtain the nonlinear recurrence characteristics of the wind fieldsignal in different sampling situations. The result provides an important basis for theselection of sensor in gas pursuit system.