UWB Impulse Radar for Through-The-Wall Imaging

Transcription

1 Internatonal Journal of Electromagnetcs and Applcatons ; (): 9-3 DOI:.593/j.jea..5 UWB Impulse Radar for Through-The-Wall Imagng P Kshore Kumar *, T Kshore kumar Department of Electroncs & Communcaton Engneerng, Natonal Insttute of Technology, Warangal, 564, Inda Abstract Ultra-Wdeband (UWB) s the promsng technology for localzaton of the objects behnd the walls. Recent terrorst actvtes and law-enforcement stuatons underscore the need for effectve through-wall detecton. UWB radar sgnals has extremely large frequency spectrum and snce low frequences has more penetraton capabltes through delectrc materals t s best sutable for Through-the-Wall Radar Imagng (TWRI). Sgnal processng n TWRI has a greater mpact n gettng the nformaton of the scanned area. Ths paper uses mpulse sgnals n TWRI, examnes the factors mpactng n TWRI and obtans the two dmensonal nformaton of the scanned scene. Electromagnetc smulaton software s used to generate the room lke structure, and to obtan the raw radar data. Keywords Ultra-Wdeband, Impulse Sgnal, Through-the-wall, Synthetc Aperture. Introducton The feld of remote sensng has developed a range of nterestng magng approaches for a varety of applcatons. Through-the-wall and through-the-buldng sensng are relatvely new areas that address the desre to see nsde structures to determne the layout of buldngs, where occupants may be, and even dentfy materals wthn the buldng. Through-wall sensng s hghly desred by polce, fre and rescue, emergency relef workers, and mltary operatons. Accurate sensng and magng can allow a polce force to obtan an accurate descrpton of a buldng n a hostage crss, or allow fre fghters to locate people trapped nsde a burnng buldng. The goals of throughthe-wall sensng technology are to provde vson nto otherwse obscured areas. Each remote-sensng applcaton area has drven dfferent sensng modaltes and magng algorthm development based upon propagaton characterstcs, sensor postonng, and safety ssues. Tradtonal optcal, radar, and sonar mage processng all begn wth basc wave physcs equatons to provde focusng to ndvdual ponts. In many radar applcatons, for example, data sampled from many sensors are mathematcally ntegrated to provde equvalent focusng usng free-space propagaton assumptons. Freespace magng s commonly seen n synthetc aperture radar (SAR) technques[] snce atmospherc dstortons are often neglgble and can be safely gnored n frst-order calculatons. Free-space propagaton does not apply for many nterest- * Correspondng author: pulkshorekumar@gmal.com (P Kshore Kumar) Publshed onlne at Copyrght Scentfc & Academc Publshng. All Rghts Reserved ng applcatons lke geophyscal sensng, medcal magng, and more recently through buldng sensng where transmsson through scatterng meda s encountered. In these applcatons, propagated sgnals dffract through a volume. As examples, geophyscal magng technques generally measure sesmc propagaton through the earth to look for dscontnutes that are often ndcators of ol, gas, water, or mneral deposts. In medcal magng, ultrasound tomography approaches account for propagaton through dfferent tssue classes[]. Non-free-space scatterng applcatons are more representtatve of the through-buldng sensng problem, albet each has ts own dstnct challenges and approaches. In through-buldng sensng, there are many ar materal nterfaces that dramatcally change the wave front. Through buldng sensors may be located some dstance away from the structure and attenuaton s largely seen only n the buldng materals and contents rather than n the large volumes of ar that occupy most of the space wthn a buldng. Sensor standoff provdes unque opportuntes and challenges for explotng buldng-dependent features. The paper s organzed as follows. Secton II descrbes the measurement scenaro used n ths paper and ts specfcatons. In secton III the effect of crosstalk and rage gatng n TWRI s descrbed. Back Projecton algorthm s explaned n secton IV. Secton V s devoted to results and Secton VI for concluson.. Measurement Scenaro A smplfed sketch of the measurement layout s shown n fg. The radar system was horzontally scanned parallel to the front brck wall at a dstance of cm from the wall as shown n fgure. The wall s consdered as the brck wall and the specfcatons are gven n table.

2 P Kshore Kumar et al.: UWB Impulse Radar for Through-The-Wall Imagng The area that s scanned behnd the wall covers the volume of 8 cm 7 cm 5 cm ndcates length, wdth and heght respectvely. The transmttng and the recevng antennas are placed at the same poston, keepng n vew to make use of the same antenna for both transmttng and recevng for further studes. hgh resoluton mages, we use an mpulse sgnals partcularly Gaussan dervatve s used. Table. Electrcal Parameters of the brck wall. Type of wall Brck wall Relatve Permttvty 4.44 Conductvty. S/m Thckness of the wall.5 m The objects that were placed behnd the wall conssts a wooden table on whch an ron cylnder s placed and at other poston a glass cylnder s placed. The specfcatons of the wooden table, ron cylnder and glass cylnder are gven n table. Table (a). Wooden table Propertes. Fgure. Sde vew of the smulated measurement scenaro. Table (b). Glass Cylnder Propertes. The resoluton s proporton to the bandwdth of the sgnal [3,4] and s gven by cτ c () R = = B Where ΔR: Range Resoluton τ: Pulse Wdth B: Bandwdth and c = 3 8 m/sec s the velocty of lght n free space. The reason of selectng Gaussan doublet or the second dervatve of the Gaussan pulse for transmttng s that t has only one peak whch can be easly notced n the doman and also has the longest lobe n the frequency doman, creatng the large bandwdth of the sgnal. Another pulse would not ft these requrements snce t would not have both the large sngle pulse n the doman and cover a large spectrum n the frequency doman[5-7]. Fgure. Top vew of the smulated measurement scenaro. One reason of selectng these materals (wood, metal and glass) s to check how the reflecton of the sgnals takes place, and the second s these materals usually found n any rooms. Snce Ultra-Wdeband (UWB) sgnals has the better penetraton propertes and has the capablty of obtanng Fgure 3. Orthographc Vew. Snce the sgnal transmtted by the antenna s the dervatve of the nput sgnal, to make use of Gaussan doublet Gaussan frst dervatve s gven as nput sgnal to the antenna. The expressons for Gaussan pulse and ts dervatves (up to second order dervatve) are gven by

3 Internatonal Journal of Electromagnetcs and Applcatons ; (): 9-3 t f ( t) = Aexp( ) τ At t f '( t) = exp( ) τ τ A t t f ''( t) = ( ) exp( ) τ τ τ The sgnal that was used here has the pulse wdth of.5 n sec and the wth centre frequency of.83 GHz and 3.3 GHz bandwdth. The plot of the Gaussan dervatve and ts spectrum s shown n fgure 3. () 3. Effect of Antenna Crosstalk and Range Gatng n TWRI Crosstalk s the part of the sgnal that travels drectly between the transmttng and recevng antennas. It s the frst and often the largest peak n the A-scan sgnal whch affects the mean value of the sgnal. Ths sgnal constantly presents n all the measurements and contans no nformaton of the scanned scene. Somes ths largest peak sgnal may cross the maxmum current ratng of the recever crcut over whch the performance may be degraded 8. Hence t s very mportant to remove the crosstalk. The crosstalk can be obtaned by measurng wth radar n the free space, or wth absorbers around (anechoc chamber). A pcr (n) = A p (n )- c p (n) (3) Where A p (n) and c p (n) are the A-scan sgnal and crosstalk at p th poston of an antenna respectvely and A pcr (n) represents the A-scan sgnal at p th poston wth crosstalk removed Crosstalk A p(n) A-Scan Sgnal at pth poston (b) Receved sgnal at a poston wth Crosstalk Fgure 4. (a) Gaussan Dervatve (b) Its frequency Spectrum A pcr(n) A-Scan Sgnal at pth poston (c) Receved sgnal at a poston wth Crosstalk Removal Fgure 5. Horn antennas were used for transmsson and recepton. Selectng horn antenna enhances the gan n the drecton of scannng and to not to consderng the regon behnd the antenna whch s of no use. The horn antenna radaton pattern s shown n fgure 5. Fgure 6. (a) Crosstalk Measurement, (b) Receved A-scan sgnal (c) A-scan sgnal after removng the crosstalk. Range Gatng s the basc preprocessng technque that decdes to fx the maxmum range of the scanned area. The recever s made to swtch on up to the delay that was fxed by the maxmum range. After that delay the sgnal s assumed to be of even farther dstance and hence gnored. The delay that was obtaned for 5cm range s.466-9

4 P Kshore Kumar et al.: UWB Impulse Radar for Through-The-Wall Imagng sec. Snce the velocty nsde the wall s gets reduced by a factor of µ rε the delay was ncreased to.55-9 sec r whch occurs at the ndex of th sample (a) the attenuaton of the sgnal. Smlarly a term should be added n eq (5) representng the reducton of velocty nsde the wall. Snce the ntenton s to obtan the mage, these terms are gven lttle mportance n ths paper. The pont spread response (PSR) whch represents the spreadng of sgnal energy n two dmensonal space of a gven A-scan sgnal s gven by f p ( x, y j ) = Ap ( t + τ ( x, y j )) (6) where =,,,3 M, j=,,,3 N, x, y j represents the pxel co-ordnates and τ ( x, yj) s the focusng delay[9] n the mage space. The PSR at p th recever poston s shown n fgure (b) Fgure 7. (a) A-scan sgnal wthout Range gatng. (b) A-scan sgnal wthout Range gatng. 4. Back Projecton Algorthm The headng of the References secton must not be numbered. Ths algorthm retreves the mage of the scanned area by makng use of the A-scan sgnals at dfferent recever postons. Snce the mage s of D and s represented by an M N matrx, the am s obtan the correspondng numercal values to fll the matrx, such that the formed mage should represent the scanned scene. Each numercal value n the M N matrx represents a pxel value of an mage. Ths mage or D sgnal s also called as B-scan sgnal[]. The numercal value at any pxel poston should be the contrbuton of the A-scan sgnals at all recever postons. Snce the range s fxed to 7 cm and the number of pxels n the range drecton s N (N represents cross range and M represents the range), the range bns are selected such that the maxmum range should not cross the prescrbed value. The A-scan sgnal receved at each recever poston assumng P targets n the regon s represented as[9,]. P A ( t) = a j= j s( t t Where A (t) s the A-scan sgnal at th recever poston, a j s the reflecton coeffcent of the j th tatget and t j s the delay of the sgnal from th transmtter poston to j th tatget poston and to th recever poston and s gven by t j R c j j ) (4) = (5) where R j s the dstance from th transmtter poston to the j th target poston. A weght factor also should be there n eq (4) representng Fgure 8. Pont Spread Response. But the fnal mage should have the contrbuton of all A-scan sgnals and hence all the PSR should be added. The expresson s gven by 5. Results B x, y ) = L p= f p ( x, y ) ( (7) Computer smulatons mplementng the mpulse sgnals for through the wall magng s done. The back projecton algorthm s used to retreve the mage of the scanned regon and the results obtaned are shown n fgure (9) and fgure (). The data collected from the recevers at dfferent recever postons s used n MATLAB to mplement the algorthm. Fgure 9. Two dmensonal mage of the scene wthout wall ndcatng the poston of the objects.

5 Internatonal Journal of Electromagnetcs and Applcatons ; (): REFERENCES [] Gorgo Franceschett, Rccardo Lanar, Synthetc Aperture Radar Processng, CRC Press, 999 [] A. C. Kak, M. Slaney, Prncples of Computrzed Tomographc Imagng, IEEE Press, New York, 988 [3] James D. Taylor, Ultra-Wde Band Radar Technology, CRC Press, The Results clearly demonstrates the poston of the objects that were there n the scanned regon. The ntersecton pont of the pont spread responses obtaned for dfferent recever poston ndcates the presence of the object. 6. Conclusons We have presented an approach of makng use of mpulse based sgnals for through the wall magng. An electromagnetc smulaton software s used to model the wall and objects nsde the room. Three dfferent materals were selected. The need for removng crosstalk and the necessasty of range gatng s also explaned. Fnally Back projecton algorthm s analyzed and used n retrevng the two dmensonal mage of the scanned regon. ACKNOWLEDGEMENTS I wsh to acknowledge my supervsor Dr. T. Kshore Kumar and Prof N. V. S. N Sarma for ther support n ths work. [4] Taylor, J. D,: Introducton to Ultra-wdeband radar systems. CRC Press, Boca Raton, FL [5] M. Ambors, Ultra-Wdeband Radar seeng Through Walls, May 7 [6] Ahmad Safaa-Jaz, Sedk M. Rad, Al Muqabel, and Ahmet Bayram, Ultra-wdeband Propagaton Measurements and Channel Modelng DARPA NETEX Program Report, November 8, [7] Sylvan Gauther and Wald Chamma, Through The Wall Survellance Techncal Memorandum, DRDC Ottawa, October [8] Mchal Aftanas Through the wall magng usng M-sequence UWB radar system a doctoral thess, Slovak. Republc. December 7 [9] Greg Barre, Ultra-wdeband Synthetc Aperture Data and mage processng Techncal Memorandum, DRDC Ottawa, January 3 [] N.T. Thanh, L. van Kempen, T.G. Savelyev, X. Zhuge, M. Aftanas, E. Zakov, M. Drutarovsky, H. Sahl, Comparson of Basc Inverson Technques for Through-Wall Imagng Usng UWB Radar, Proceedngs of 5th European Radar Conference, October 8 [] Fauza Ahmad, Moeness G. Amn, Saleem & A. Kassam, Synthetc Aperture Beamformer for Imagng Through a Delectrc Wall, IEEE Transactons on Aerospace and Electronc Systems, Vol 4 No, 5, 7-8