Multipath Signals Suppression And Exploitation Algorithms In Urban Environment

By utilizing the penetrating and multipath reflecting signal of electromagnetic(EM)waves,radar system can detect,locate,image and track the hidden targets inside buildings and behind building corners,which has important application prospects in urban street warfare,anti-terrorism rescue,public security law enforcement,disaster rescue and other fields.When EM wave propagates in the building environment,complex physical phenomena such as reflection,transmission,diffraction,attenuation and dispersion will occur,lead to the difficulty of detecting hidden targets.Acquiring the locations of the hidden targets under the condition of cluttered echoes is a research hot and difficult issue in international research.Focusing the hidden targets detection problem aformentioned,this thesis studies the multipath signals suppression and exploitation algorithms in urban environment.Initial achievements have been obtained in multipath ghost suppression,multipath signals based building layout estimation and multipath signals based non-line of sight(NLOS)target localization.The main innovations of this thesis are as follows:1.A similarity matrix based multipath suppression algorithm is proposed.By constructing the similarity matrix between these focus areas in radar image,the multipath ghosts are able to be associated with the corresponding true targets.The proposed algorithm conquers the problem of target loss caused by overlapping of true targets and multipath ghosts,enabling the effective suppression of the multipath ghosts in multiple targets scenario.2.An imaging dictionary based multipath ghost suppression method is proposed.Several different types of imaging dictionaries with respect to direct propagation path and multipaths are constructed.The similarities and differences of these radar images derived from different imaging dictionaries are analyzed.By fusing these radar images aforementioned,the high-quality multiple targets image is obtained.3.An algorithm for suppressing the multipath ghosts in through-wall radar image based on the detected angle dependence is proposed.Based on the derived position distribution characteristics of the multipath ghosts in these radar images that origined from different array detected angles,the array rotation strategy is adopted to remove the multipath ghosts from different radar images,addressing the multipath ghost suppression problem under the variable detected angel situation.4.An algorithm for wall position estimation based on multipath signals exploiting is proposed.By utilizing the geometric relationship between the back wall multipath ghosts and the true target,the positions of the inner surfaces of the front wall and the back wall are able to be estimated.Moreover,a method for building geometric layout estimation is proposed by means of tracking true targets and multipath ghosts.For the images correspond to two vertical views,the meanshift-based tracking algorithm is employed to obtain the corresponding tracks of true targets and multipath ghosts.Based on the relation between the true target track and multipath ghosts' tracks,the building layout estimation is achieved under the fixed radar system position situation.5.According to the propagation characteristic of the combined path of corner diffraction and wall reflection,a diffraction propagation delay based NLOS target localization algorithm is proposed to obtain the position of NLOS target hidden behind the corner of single building scenario.By extracting the echo delay difference between the round-trip diffraction path and the combination path of corner diffraction and wall reflection,the accurate location of the hidden target can be calculated.Moreover,an image region matching based NLOS target localization method is proposed obtain the position of NLOS target hidden behind the corner of L-shape building scenario.By matching and correlating different focus areas in radar image,the localization of the hidden target is implemented.