Estimation of moving targets located behind reinforced concrete - PowerPoint PPT Presentation

Estimation of moving targets located behind reinforced concrete wall using moving sensor arrays Marija M. Nikoli � 1 , Arye Nehorai 2 , and Antonije R. Djordjevi � 1 1 University of Belgrade, Serbia 2 Washington University in St. Louis, USA COST Action IC0603 Workshop Cyprus, April 2008 �

Contents � Estimation of personnel hidden in buildings reinforced with concrete walls � Through-wall vision � Measurements by sensor arrays moving in front of the building � Recent investigations consider homogeneous walls � Parameters of the wall: � Unknown � Known � The influence of the wall on the signal propagation taken into account � 2D electromagnetic simulation �

Forward model � We consider reinforcement with metallic bars � Simulation model: � Stationary and moving objects � Dielectric and metallic bodies � Sensors � Filament conductors � 2-D field simulator � MoM solution � Equivalent surface electric and magnetic currents � Bars modeled as PEC wires �

Simulation model � Array consists of M sensors � Measurements are taken at � N known positions � L frequencies � In the 2-D model, induced fields are used instead of electromotive forces M � � � � � � � noise E n f E n f w , , i l ik l � k 1 � The array is a multiport network M � � � � � � � � � E n f z n f I f w , , , i l ik l k l k � 1 �

Time-domain response � Signals reflected from the wall � Homogeneous � Reinforced � Sensors transmit short pulses � Reflections are separated in the time domain �

Time-domain response � Reflections from a metallic plate behind the wall � Reinforced walls do not introduce additional delays � The signal is significantly attenuated �

Estimation of wall permittivity � Signals reflected from the front side contain information about wall permittivity � We focus the measurements in the time domain with respect to the front side of the wall N M M � � � � � � ��� 1 � � � E t E n t s n , ik ik ik 0 NM 2 � � � n i k 1 1 1 �

Estimation of wall permittivity � h ( t ) - reference waveform � h ( t- 1m/ c 0 ) -waveform induced in Rx at a distance of 1m from the Tx T 1 � � � � � E t h t dt 0 � R ˆ 0 2 � � T 1 � � � � � h t dt � � � � 0 2 � � � ˆ R 1 � � � � ˆ � � r � ˆ R � � 1 �

Estimation of wall thickness � Measurements are focused with respect to the back side of the wall � � � � � � � � � � � � � � � � w y y w 2 c � � - /cos / sin / w � � r r 0 � � � � � � � � � � � � � � � � x x y y a d 2 arg min - t n sin / sin k i i w r � T N M M � � � � � � � � � � � � I w � E n t � � w h t dt , ik � � � n i k 1 1 1 0 �

Estimation of targets � Assumptions � Static scene measurements available � Walls, furniture � Dynamic scene measurements available � Static scene altered by the appearance of targets � Difference in measurements associated with fields scattered from targets � Parameters of bars need not be known ��

Estimation of targets � The space in the building is divided to small cells 1 � The scattered field is focused with respect to different cells ( x , y ) 1 F. Ahmad, M. G. Amin, and S. A. Kassam, “Synthetic aperture beamformer for imaging through a dielectric wall,” IEEE Trans. Aerosp. Electron. Syst. , vol. 41, no. 1, pp. 271–283, Jan. 2005 ��

Estimation of targets � � � � � � � � � � � � � � � � x y y y w w 2 c � � , - - /cos / sin / i k i k i k i k � � , , , r r , 0 N M M � � ��� � � � � � � � � � � � � � E t x y E t x y x y ; , , , ik i k � � � n i k 1 1 1 T � � � � � � dt � � � I x y E t x y h t , ; , 0 ��

Modeling bar distortion � Bars significantly distort waveform � Estimation is improved when distortion is modeled � New reference pulse � � � t � ~ � � � � � � � � � � � � � � � h t F H f T f f � � 2 1 rect , 0 exp j 2 T � � ��

Simulation scenario � Targets � Array � PEC cylinders, radius 0.2m � 5 sensors � Separation between sensors � Walls 0.2m � Thickness w = 0.2m � Standoff distance 0.75m � Permittivity � r = 3 � Measurements every 0.2m � Bar diameter D bar = 2cm � Bar period d bar = 15cm � Frequency � f max = 2GHz, � f = 5MHz ��

Wall parameter estimation � Estimated permittivity � r = 2.94 � There is no unique solution for wall thickness! ��

Two targets ��

Improved quality when parameters of bars are known ��

Erroneous wall thickness ��

Conclusions � Reinforced wall significantly attenuates low frequencies and distorts transmitted waveforms � Signals reflected from the objects behind the wall have an oscillating nature � Scene images are blurred because of the multiple echoes � We assumed we have available � Measurements of the static scene behind the wall � Measurements altered by the appearance of people ��

Conclusions � Measurements are coherently added with respect to different time-delays to estimate � Wall permittivity � Wall thickness � Position and number of targets � The algorithm performed satisfactorily when bar parameters are unknown � If the bar parameters are known: � Estimation is significantly improved � The minimal necessary SNR is lower compared to the case where the influence of the bars on signal shape is ignored ��