Application Of HHT For Noise Suppression In Radar Echo Signal

UWB bioradar is widely applied in the field of disaster rescue and thefight against terrorism for its prominent advantages such as range resolutiongood penetration and anti-disturbs.In practice, before the electromagnetic wavesemitted by bioradar reach human target, it need penetrate some media like ruinor obstacle. So the radar echo signal usually contain much noise which will haveinfluence on the detection of vital signs. In order to draw the vital signs in radarecho signal, the way of noise suppression is absolutely necessary.In previous research, we employed FFT to reduce high-frequency noiseand reserve low-frequency signal. But the vital signs detected by bioradar arenon-stationary which don’t meet the stationary assumption for FFT. And FFTbelongs to global transformation, in the process of noise suppression, mutationposition is also removed which may carry important information. To solve theseproblem, a new way of noise suppression based on HHT is introduced. UWBradar echo signal is decomposed into IMF by EMD, our work is to find thecut-off point of IMF between two different parts in which noise and usefulsignal take the dominant position respectively. What we need do next is toextract the useful information of those IMFs in which noise take the dominant position and combine it with the IMFs in which useful signal is dominant.This thesis focuses on three aspects as follow1A deep study has been carried on HHT and the meaning of someimportant notions in HHT have been analysed, such as instantaneous frequency,intrinsic mode function and the spectrum of Hilbert, the advantages anddisadvantages have also been discussed.2Completing the algorithm of EMD which is the key of HHT, after UWBradar echo signal is decomposed into IMF by EMD, we study the energydistribution of IMF, and two different way of noise suppression are takenaccording to the condition whether there is a minimum of IMF’ energy or not.The results show that these two different ways can both reduce high-frequencynoise and keep low-frequency useful signal effectively.3Comparing the results of noise suppression based on HHT and0.5Hzlow pass filter, we can find the former not only reduce high-frequency noise, butalso reserve the details in mutation position.Innovation in this thesis are listed below:1Employing HHT in the process of UWB bioradar echo signal, andapplying it in the collection of signal of human breath for the first time.2According to the two different conditions that the SNR of radar echosignal is high or low, we employ the way based on the analysis of the IMF’energy and the way based on the different feature of ACF of noise and signalrespectively, and they both work effectively. In order to reserve the details insignal completely a way which is similar to wavelet soft-threshold de-noising isemployed to draw the useful information in those IMF where noise take the dominant position.