o ~ ~ ~ Cr -~-- - - ~'-.< -~ 4.'.-- '--'C'. O4-7gts~, ~. ;n "4) ~ ~ 4~-. r~-~- A * 4 \$4-~4$4<--'4a,.AN.{ t4&..t~- '~ -, 7~~-'~ A -Wt.wt ~A~9o~kVFA ~ -'~~'~ -I ---~'r{ ~ C-" o-~- -r ; r~y-td~.c -o - -'V"- Z ~.r.x - -. -& -i -±~.t. _ - ~rj -"-'-4-i "~ -4 - '~A --. 4'., -- 4-A' ~""o-'- -o -4 -- ~- ~tt-~z~~ 'C-F ~ - ~- -* P "S - 4' '1 I! H - -4 ', A" - ' 7" {~ %y~u4t - 10;; -l -Y '4" S U; M~c;Th~ Lr'. Ft F', ftp - - - -. rbkt' ON' Lt 2Q~A7c2r ~- " - Cl'N. I.'' - - Ur4- v ~ s 4 ct 4Iur.A~cs - '-'>4's F'' - '- 1' F' F.-A2fl't.~tF~~~~'1 B t 4 ~ 4-t4 ~ 'F" Uvtc docv;wutt 1 tn~ 'beew 4~9tt'Y~d Lot ~x~iic xeliw4; c~j COlt. ~43 tfv"ttttl>x4yf'jc-' J~ 44 "'k'".' 4 '" F'~"" 4 "''"-"-'-44 F> ~5',?A4, t A A" -~ - k4- 't>-f. F' '.> '~~~A -. F F-.~. >'~""-' C- *>'jsy~,'. 4 V - '.Xs~'--:'i'-'~ F 74- '>7' '-A 4 g -
Best Available Copy
t No 0. 4 73 'a A... i tv Classification 1 / 2 DOCUMNT CONTROL DATA. &D 'Scurit Clnssilication of title. body of obserect snj arem on.notari.vi -pubt br entered Pien rho esverail fepatt to eaaalio~d) 1. ORK(. RATIGAT ry C,a.uhw 1.~A. CCAT SECUREITY CLASSIFICATION ENGINE"ERINGiAI TIAL R EXPERIMENT STATION UNIVERSITY OF FLORIDA GAINESVILLE, FLORIDA 32601 Js REPOPT TITLE q.4uncasfe RANGE LAWS FOR.DISTANCE MEASURING SYSTEMS USING FREQUENCY MODULATION WITH A NONLINIEAR TRIANGULAR WAVESHAPE _A F. O -SC-c P rv C NOTES (?7'po of report and inclumov. dares) S. AU TsORIS) (First nme* a,,td*e initial. 1* last) Leon W. Couch * 6. REPORT DATE 7a. TOTAL No or PAGES lit.. NO.o CVrEs March 28, 1972 54 60 90. CONT64ACT 0011 GRANT NO o.i"a io0s PIEVoRT 9.t*NE"qI DAAG39-70-C-0058 b. noirc T No. 0058-8 I{DL Pro.j: 16114.,ga.. tumb.udl aboaefrd DA-1B262301A208 IAto N.P.We) SR-72-3 damcms Code: 522A.11.17700 10 OISTRIRUT40ON STATEMER? This document has been approved for public release and sale;4 its distribution is unlimited. ARMY MATERIEL COMMAND HARRY DIAMOND LABORATORIES WASHINGTON,D.C._20438 Range laws are obtained for a distance measuring system which transmits an RF signal. The RF is frequency modulated by a nonlinear triangular waveform, the amount of nonlinearity being specified by a parameter.ri The nonlinearity may have desirable or undesirable effects; that is, the sidelobe level of the range response may be decreased or increased depending on the shape and amount of nonlinearity. The results of this report may be used to predict range laws of system which have nonlinear triangular modulation. Alternately, practical nonlinear triangular waveforms may be specified which will reduce the range-law sidelobes. DOPOSI I4 cs S-rcrstv riali
-4 Security Ctassilication rlink A LINKS8 LINK C OLO WT ROLE WT IROLE IV 1. Frequency Modulation 9 0 2. Nonlinear Triangular Frequency Modulation 8 3 3. Radar 1 0 4. Range Laws 9 3I 5. Sidelobe Suppression 8 3 6. Triangle Waveform Modulation 9 secutityclasfrilcation
DA-1B262301A208 AMCMS Code: 522A.11.17700 HDL Proj: 16114 AD 0058-3 RANGE LAWS FOR DISTANCE MEASURING SYSTE!S USING FREQUENCY MODULATION WITH A NONLINEAR TRIANGULAR WAVESHAPE by L. W. Couch March 28, 1972 United States Army Materiel Command HARRY DIAMOND LABORATORIES Washington, D. C. 20438 Prepared by: ENGINEERING AND INDUSTRIAL EXPERIMENT STATION ii College of Engineering University of Florida Gainesville, Florida 32601 Under Contract Number DAAG39-70-C-0058 This document has been approved for public release ana sale; its distribution is unlimited.
ABSTRACT Range laws are obtained for a distance measuring system which transmits an RF signal. The RF is frequency modulated by a nonlinear triangular waveform, the amount of nonlinearity being specified by a parameter. j The nonlinearity may have desirable or undesirable effects; that is, the sidelobe level of the range response may be decreased or increased depending on the shape and amount of nonlinearity. The results of this report may be used to predict range laws of systems which have nonlinear triangular modulation. Alternately, practical nonlinear triangular waveforms may be specified which will reduce the range-law sidelobes.- I Preceding page blank 3
TABLE OF CONTEN,'TS Page No. ABSTRACT...... 3 KEY TO SYMBOLS...... 7 1. INTRODUCTION...... 9 I 2.2. 2. GENERAL FORMULATION FOR THE RANGE LAW...... 9 2.1. Range Law in Terms of the Fourier Coefficients..... 9 Relationship Between the Fou-ier Coefficients and the Ambiguity Function...... 12 3. RANGE LAW FOR TRIANGLE PLUS SINUSOIDAL NONLINEARITY...14 4. RANGE LAW FOR TRIANGLE PLUS CUBE-LAW NONLINEARITY.... 27 5. EXPERIMENTAL AND THLORETICAL VERIFICATION OF RESULTS..... 35 5.1. Experimental Verification....... 35 5.2. Theoretical Verification......... 38 6. SUMMARY........ 41 APPENDIX A--Fourier Series for the Hixer Output....... 43 APPENDIX B--Fourier Coefficients for Triangle plus Sinusoidal Modulation...... 46 APPENDIX C--Fourier Coefficients for Triangle plus Cube-Law Modulation....... 48 APPENDIX D--Percentage of Modulation Required for Equivalent Sinusoidal and Cube-Law Nonlinearity....... 50 REFERENCES........ 51 DISTRIBUTION...... 52 LIST OF FIGUIRES 1. DMS and a Point Target 0... 2. Triangle pius Sinusoi1l N1 onlinearity...... 14 3. Range Response for K = 0, Sinusoidal Nonlinearity....... 17 4. Range Response for K = 1, Sinusoidal Nonlinearity...... 17 5. Range Response for K = 2, Sinusoidal Nonlinearity....... 18 6. Range Respons, for K = 3, Sinusoidal Nonlinearity...... 18 7. Range Response for K 4, Sinusoidal Nonlinearity...... 19 8. Range Response for K = 8, Sinusoidal Nonlinearitv...... 20 Preceding page blank 5
LIST OF FIGURES Page No. 9. Range Response for K = 16, Sinusoidal Nonlinearity...... 21 10. Range Response for K = 0, Negative Sine Nonlinearity..... 22 11. Range Response for K - 1, Negative Sine Nonlinearity..... 22 12. Range Response for K = 2, Negative Sine Nonlinearity... 23 13. Range Response for K = 3, Negative Sine Nonlinearity..... 23 14. Range Response for K - 4, Negative Sine Nonlinearity... 24 15. Range Response for K - 8, Negative Sine Nonlinearity... 25 16. Range Response for K = 16, Negative Sine Nonlinearity.... 26 17. Triangle plus Cube-Law Nonlinearity....... 27 18. Range Response for K = 0, Cube-Law Nonlinearity...... 30 19. Range Response for K = 1, Cube-Law Nonlinearity....... 30 20. Range Response for K - 2, Cube-Law Nonlinearity. &.... 31 21. Range Response for K = 3, Cube-Law Nonlinearity..... 31 22. Range Response for K - 4, Cube-Law Nonlinearity....... 32 23. Range Response for K = 8, Cube-Law Nonlinearity....... 33 24. Range Respense for K - 16, Cube-Law Nonlinearity....... 34 25. Test Apparatus for Experimental Measurement of Range Laws 35 26. Experimental Range Laws for K = 4, Triangle plus Positive Cube-Law Nonlinearity....... 36 27. Experimental Range Laws for K = 8, Tridngie plus Positive Cube-Law Nonlinearity..... 37 28. Pc vs I. for Approximately Equivalent Cube-Law and ~~Sinusoil'al Nonlinearities.......39 29. Comparison of Range Laws, Cube-Law Nonlinearity K = 4.... 40 30. Nonlinear Frequency Modulating Wavefo,:ms..... 41 6
KEY TO SYMBOLS B ck D DMS m(t) loop c loops loopt Rk(-r) Peak-to-peak FM deviation (Hz) Fourier coefficient of the mixer output for the kth harmonic of the fundamental modulation frequency Deviation sensitivity of the FM modulation (radians/volt-sec) Distance measuring system Composite frequency modulating waveform Percent cube-law nonlinearity Percent sinusoidal nonlinearity Percent of pure triangular contribution to the composite nonlinear trianpular modulating waveform Range law associated jith the kth harmonic of the fundamental modulation frequency v(t) = e 0 Period of the modulation t j f m(t)d-r Complex envelope of the transmitted RF signal x 2BT Normalized round-trip t,, delay between the distance measuringdevice and the target x 0 = 2BT 0 y(t,t) Od(t,T) Tr c = 2lTfc Cod = 27fd Normalized round-trip time delay which gives the peak range law response Mixer output signal Difference phase between the transmitted and received RF signals Round-trip time delay between the distance-measuring,-device and the target Radian carrier frequency Radian doppler frequency dod(t,t") Wid(tT) = dt Instantaneous difference frequency (ra.'stc) between the transmitted and received RF signals wm = 27fm -- Fundamental modulating frequency (rad/sec) IT
RANGE LAWS FOR DISTANCE MEASURING SYSTEMS USING FREQUENCY MODULATION WITH A NONLINEAR TRIANGULAR WAVESHAPE 1. INTRODUCTION The range law of a distance-measuring system (DMS) is the envelope of the test statistic [1]. This is a function of the time delay between the device and the target (or the corresponding distance between the device and the target). If this envelope is above a certain level determined by a pre-set threshold,the device is said to be within the functioning window of the target and an alarm signal is given at the output of the threshold device. Here the DMS is assumed to be of the type that is frequency modulated by a periodic waveform. It is well-known that nonlinear frequency modulation can be used to decrease the sidelobe levels in the range law [2, Chapter 20],[6]. In fact, by frequency modulating with a waveform having Taylor weighing, the sidelobe may be decreased to 40dB below the main lobe, as compared to 13dB sidelobe suppression for pure triangular modulation [2]. The modulating waveform for Taylor weighing is 7 m(t) = a + Kn sin (nimt Itl < T where T is the period of the modulation and wn = 2n/T. For the DC range-law to have 40dB sidelobe suppression, the constants are K 1 = -0.1145, K 2 = 0.0396, K, = -0.0202, K 4 = 0.0118, K 5 = -0.0082, K 6 = 0.0055 and K 7 = -0.0040 [2]. In practice this modulating waveform is difficult to generate. In this report the range laws for the DMS will be calculated for a class of nonlinear frequency modulating waveforms. This class consists of (I) Triangle plus one sinusoidal term and(2) Triangle plus a cube-law term. This class of waveforms is relatively easy to generate in practice. Thus, the results of this report can be used to design practical nonlinear modulating waveforms which will increase the sidelobe suppression over that obtained by use of pure triangular modulation. Alternately, this class of waveforms can be used to model the actual nonlinear modulation in devices which are designed to have pure triangular modulation. Thus, the realized sidelobe suppression of an actual DMS with "triangular" modulation can be predicted. This is important because, for certain nonlinear shapes, the sidelobe suppression is increased (instead of decreased) by the incidental nonlinearity. 2. GENERAL FORMULATION FOR THE RANGE LAW 2.1 Range Law in Terms of the Fourier Coefficients For periodic modulation the range law is obtained by evaluating the Fourier coefficient for the mixer output of the DMS. Referring, to Figure 1, Preceding page blank
x(t) DMS ~r(t) 1re Mixer Output y(t,t) Figure 1. DI4S and a Point Target the transmitted signal is represented by x(t) = /r Re {v(t)ejwct} (1) where wc is the radian carrier frequency and v(t) is the complex envelope [3]. Then the corresponding received signal is j t r(t) - Re {v(t--)ej~c(t - )} (2) where T is the round-trip time delay. To the first approximation, the mixer output is given by the cross-product term y(t,t) = x(t)r(t) = Re fv(t)v*(t-)e:jw C T + Re fv(t)v(t-t)e j ia c ( 2 t - T ) where [.]* denotes the comple; coniugate of [.]. Neglecting the output about 2 wc (which is tiltered out in the ph-sical D'!S), we have for the mixer output y(t,t) = Re {v(t)v*(t-t)ejw c CT (3) The test statistic is a filtered version of (3) usually obtained by a doppler band-pass filter center'd about some harmonic of the fundamental modulating frequency. For this frequency modulated DMS II)fm(t)dt] v ( t ) = e 0 f,: 10
where D is the deviation sensitivity (radians/volt-sec) and m(t) is the periodic frequency modulating waveform. Then where iod(t,tr) v(t)v*(t-t) = e (4) Od(t,T) = D t t-t m(o)do Od(t,T) is the mixer output difference phase with T being a parameter. The instantaneous difference radian frequency-associated witii the mixer output is then WiddttT) D[m(t) - M(t-T)] (5) where it is assumed that the variation Of T with respect to t is negligible, i.e. the doppler frequency is much less than the carrier frequency. Thus, 0d (t~t D f y a(t) - m(t-t)]d-r (6) If m(t) is periodic, then, Od(t,") is periodic with respect to t. Thus, from (4), v(t)v*(t-r) is also periodic and can be expanded in a Fourier series v(t)v*(t-t) = ejad(t = 'T) = k- c ke~ kwmt k m (7) k=-oo where Tj 0 T e j 0 d(tt) e-jk(% t dt8 (8) Wm= 27r T and T is the period of the modu'ating waveform. Using the above formulation -he mixer output can I), obtaincd in terms of the Fourier coefficients {Ck}. Assuming a modulating wavef,-,m such that the instantaneous difference frequency has half-wave odd symmetry; that is Wil(t,T) = -id(t-t1
f- -- -- then from Appendix A, it follows that the mixer output is y(t,t) = Ico(T)I cos [wd(t-to) + C I 00- where + I 21ck(T) cos Ewd(t-to) + &0 cos [kwm(t-to]] 9) k=l to is the value of t at the first turn-around time of the modulation m(t), I [.]I denotes the magnitude of [] and Al denotes the angle of [.] The range law is given by the envelope of the output of the doppler-band filter centered at w = kw m. Thus, the range ]a,,,, are Rk(T) = L F c(t)j, k = 0I (10) 21Ck(T)l, k --- L, t. where ck(t ) T f ej e d ( t T) e - J k w mt dt.(1,.- 0 2.2 Relationship Between the Fourier Coefficients and the Ambiguity Function It is interesting to relate ck(t) with the Ambiguity Function. The Ambiguity Function is given by [1, p. 353] X(T,W) f v(t - T)v*(t +!!)e - dt which becomes for periodic complex envelopes X(T,W) f v(t - 'YT)v*(t + T)e - J wt dt. (12) I0Using (7), (11) and (12) we obtain the relationship WMT C(T) = (-T,km) (13) 12
I or I!Ck () = A(-T,kwm)I (14) Thus, the magnitude of the Fourier coefflcient gives the magnitude of the Ambiguity Function along the w = kwm axis where the positive and negative directions of the t axis have been reversed. Furthermore, the Fourier coefficient corresponding to the DC range law, co(t), is related to the power spectrum of the transmitted signal by the Fourier transform. From (12) for the case of w = 0, A(,0) is the autocorrelation, Rv(T), of the complex envelope v(t) of the transmitted RF signal, x(t). That is, for w 0, (13) becomes i Rv(T = c,;(-t) = X(T,O) and P*(i) = Rv(-T). The power spectrum of the transmitted signal is ginen by [3] Px(W) = [rv(w-wo) + Pv(-Wr'wo)] Pv(w) and Rv(r) are a Fourier transform pair. Thus C(T) =X*(TO) f Pv(w)e " - ' ) dw 2-r or B/2 CO(T* ) = f Px (UIWc)e - j WT dw (15) 0 B 1 2x wh-ere Px(wa) is the power spectrum of the transmitted RF signal, x(t). B is the bandwidth (rad/sec) of the RF signal Wc is the carrier frequency (rad/sec). 13
Thus, the c 0 (T) coefficient, corresponding to the DC range law and the conjugate of the Ambiguity Function evaluated along the w = 0 axis are related to the translated power spectrum of the RF signal by the Fourier transform times 2/w. A detailed discussion of the Ambiguity Function and its application to DHS is given by Rihaczek [4]. 3. RANGE LAW FOR TRIANGLE PLUS SINUSOIDAL NONLINEARITY The range laws, Rk(T), will now be evaluated assuming periodic modulation by a composite waveform which is the sum of a pure triangle plus a sinusoid as shown in Figure 2 below. The composite modulation is described by where m(t) = mi(t) - b cos wmt (16) ra(t T/4) 0 < t <_ T/2 m 1 (t) ] (17) La(t 3/4 T),T/2 :S t < T a and b are constants to be determined according to the percentage of sinusoidal nonlinearity. M, (t)l T/ t (a) Triangular Component M2(t) / I / t (b) Stnusoidal Component me(t) (c) Composite Modulation Waveform Figure 2. Triangle Plus Sinusoldal Nonlinearity 14
The percentage of sinusoidal nonlinearity added to the pure triangle component is defined by the percentage of sinusoidal contribution to the cmwposite peak value measured at the tine that the peak-value occurs. Thus, the percent sinusoidal nonlinearity is given by 100P s where A b = b + at/4 (18) The percentage of modulation, see (18), may be positive or negative depending oh-the sign of b with respect to a. It will be found that the sidelobes of the range law are increased or decreased depending on the sign of Ps. The range laws resulting from this triangle plus sinusoidal nonlinearity are described by (see Appendix B for the derivation) where Ico(T)I k =0 Rk(r) = (19) 2jCk(T)I, k =1,2. k 1 ej(lr 2 Ps y("! Psx), O-xPt)] + (_1)ke j Ptx Y[(- Ps x ), (k + xpt)] (20) and Y[z,vl 0 f e-j[- cos e + vo] do (21) A (22) x =2Br 22 Pt 1 I - PS (23) and B is the peak-to-peak frequency deviation (l1z) of the composite frequency modulating waveform. For B >> fm, which is usually the case, B corresponds to the spectral bandwidth (Hz) of the transmitted RF signal. to For the case of Ps = 0, i.e., pure triangular modulation, (20) reduces i (x-k) sin [I (x-k)] sin [1 (X+k) c = k e 7r + + (-)(- (24) 2 (x-k).x+k) j which yields the well-known [5] sin z/z type of range law. 1!
For the case Ps = 1.0, i.e. pure sinusoidal modulation, (20) reduces to Ck = (- 1 )k e 3 k G' x) (25) which yields the well-known [2] Bessel function range law. The range laws for several percentages of sinusoidal nonlinearity are shown in Figures 3-16. Tht- e curves were obtained by evaluating (20) numerically. Figures 3-9 give the range laws for positive percentage of modulation. It is evident that this type of nonlinearity moves the range laws to the left; that is, it decreases the normalized delay, xg, to the peak of the range law. This positive percentage of sinusoidal nonlinearity also increases the sidelobe level over that obtained for pure triangle modulation for distances larger than x 0. This type cf nonlinearity is undesirable. Figures 10-16 give the range laws for negative percentages of modulation. This type of nonlinearity moves the range laws to the right, i.e. increases xd. However, the sidelobe level is decreased for distances greater than x 0. From Zhese figures, the optimm decrease in the sidelobe level is obtained when the percentage of siausoidal nonlinearity is set to approximately This type of l6earily is desirable. Ps = -1.5 (1/k+4) (26) 16
: *O Sn S. l02 Sind +0. 100% T~angle:.. 0 6....... 50% Sine C) 100% Sine 0.4- O.ZTrange 25% Sine 100%Sin '00 Sine 0% Tri- 0.2 angle - 0 x3 1 4 5 6 7 x 2B Figure 3. Range fesponse for K = 0 Sinusoidal Nonlinearitv 0[ Sine ; 100% Triangle,102 Sine / 25% Sine \ io50 Sine ' "0.6 0.4 100% Sine 0.2 0% Triangle I J I 0 1 2 3 4 5 7 Figure 4. Range Response for K = 1 Sinusoidal Nonlinearity x - 2BT
1. 0% 0-Sine -- - 100 Triagle 0. 100 Sine -v 5% Sine 04 \ 50% Sine ;.- 0. 2-100% 0% Triangle Sine 21J2( y X), A 0 1 2 3 4 5 6 7 8 9". x =2BTx Figure 5. Range Response for K = 2 Sinusoidal Nonlinearity 1.0..... 0% Sine - - - 100% Triangle t 0.8 l Sine ~. 5% Sine C14 50% Sine 0.4-- 02100% Sine / 0.2-- 0% Triang 2 JIT,-- x) I 0 1 2 3 4 5 6 7 8 91 x 2BT 4 Figure 6. Range Response for K 3 Sinusoidal Nonlinearity
-I. 0% Sine 1.0 100% Triangle 1 0.8 1/ II.10.6 50 Sine- I-.4. ~~ J4 ~ (, 'X)3:. 0.4--'7N 2 2 100% Sine Triangle 0.2~ 7Ti0n :0. 2 I1/.jx) / e 2 3 4 5 6 7 8 9 10 11! x = 2BT Figure 7. Range Response for K 4 Sinusoidal Nonlinearity 19
II C% ri U D 00 0 0 14) :0 0 000 C - 00 a,- -4 0 0 0 1 9 0jZ = ())
C3 0 or X 0~r HG Is c r.- -N Z6 1-4- 4' 1-24 ua. 0 * ai 10 ca' -H $4 '.4a 0% 000
- " - a I. 8...... I o a-,- i :... i 0.6I *... - - I :... 0.-4 125Z sin "" VOI Triangle. I V Figure 10. Range Response for K =0 Negative Sine Nonlinearity 2=S. - I 0.0 I....... * 00%~ Triangle / a ii. 0.4-10%. - I 1 0.2 0% Sine in 0 1 - L L J.. :.I AT Figure 1. Range Response for K = 1 Vegative Sine onlinearity 22
---------------- - -T r '~~ ~ ---- I... : -25' Sine r 0.4 i -IOt Sine i U 1.00 Figure 12. Fange Response for K =2. Negative Sine Nonlinearity o - 2 si- / e /' ' p,0.8/ 107 Sine 0.6. -25% Sine 0.4-- -50 Sine 0O- Sine 0.2 100 Triangle- 7T- 0.2- IM". Triangle 0 1 2 3 4 58 Figure 13. Range Response for K 3 Negative Sine Nonlinearity 23
-0.8 -- 102' Sine 0.6.- 25% Sim e /500-0.4---0 0SiSin SSiee *0.2-2 3 4 5 6 7 8 9 10 1 x 2BTY Figure 14. Range Response for K 4 Negative Sine Nonlinearity
-41 Go~ a z *Lr I i 0) *- - - I m : Iaiv4 a, * '8-00 TI -. I Oca U) - 0) -. *~-V4-- ~ I~S *J -4.- 04 E-4- I a co - Z --- a -- -0.--- a~25 -,
I In I I n I-..- T------.i--.-. > r- (3 - -4 4-4, E-4 a~ NO 0 0 0 l C. 26
4. RAWNE LAM FOR TRIANLE PLUS CUBE-LAW.ONLINFARITY The rai:se lays, kk(t), wili now be evaluated assuming periodic modulation by a composite wveform which is the sum of a pure triangle plus a cabe-lw aw shown in Figure 17 below. %z) Triangular Component U2 (t)[ (b) Cube-Law Component m(t) t (c) Composite Modulation Waveform Figure 17. Triangle Plus Cube-Law Nonlinearity The composite waveform is described by where wm(t) = mi(t) + m 2 (t) (27) -c(t T/4),0 < t T/2 r I (t) (28) -c(t (3/4) T),T/2 <t T? 27
and -d ( -d(t T) ( )/2sts c and d are constants to be determined by the percentage of cube-law nonlinearity. Just as in Section 3, the amount of cube-law nonlinearity added to the triangular component is defined by the percentage of cube-law contribution to the composite peak value measured at the time that the peak value occurs. Thus, the percent cube-law nonlinearity is given by loopc where PC d(t/4) 3 d(t/4) 2 (30) Pc = =. (30) d(t/4) 3 + c(t/4) d(t/4) 2 + c The range laws resulting from this triangle plus cube-law nonlinearity are described by (see Appendix C for the derivation) F Ico(T)I, k = 0 Rk(T = (31) R~k~r) = 2LCk(T)l, k = 1,2 j3 where I J{Pte + PC(2/7r)2[(G-Tr/2)3 + (Tr/2)3])x Ck() r e cos(ko)do (32) x = 2B (33) Pt 1 Pc (34) and B is the peak-to-peak frequency deviation (12z) of the composite frequency modulating waveform. For B >> fm, which is usually the case, B corresponds to the spectral bandwidth (Hz) of the transmitted RF signal. For the case of Pc = 0, i.e. pure triangular modulation, (32) reduces to our well-known sin z/z type result given by (24). Once again the percentage of modulation may be positive or negative depending on the sign of d with respect to c in (30). By comparing the shapes of the sinusoidal and cube-law modulating waveforms (see Figure 2 and 17) it is seen that the shape obtained for a positive cube-law nonlinearity is roughly equivalent to that obtained using a negative sinusoidal nonlinearity. (The exact relationship is discussed in Section 5.2). Thus, in comparing the results 28
for cube-law nonlinearity with those for sinusoidal nonlinearity, we would expect the sidelobe level in the range laws obtained for small percentages of Positive cube-law nonlinearity to be suppressed with respect with those obtained usivg pure triangular modulation. The converse is also true. Since only the positive cube-law nonlinearity gives sidelobe suppression while negative cube-law nonlinearity gives increases in the sidelobe levels, only the positive cube-law results will be given here. The range laws for several positive percentages of cube-law nonlinearity are shown in Figures 18-24. These curves were obtained by evaluating (32) numerically. It is evident that this type of nonlinearity moves the range laws to the right; that is, it increases x 0. Furthermore, if the percent nonlinearity is not too large the sidelobe level is decreased at distances greater than x 0. From these figures, the optimum decrease in sidelobe level is obtained when the (positive) percentage of cube-law nonlinearity is set to approximately Pc l 1/k+4, (35) Thus, some positive cube-law noiinearity is desirablp, and negative cube-law nonlinearity is undesirable. ' ii Ii I 29
i - t.8 --, 0Z _,....-. :... I ;.... -O~t-, ----........ ' 10lO Cube 1.,4 - " -0.2 - ' i 100% O0% Cdbe Triangle i 0 3 4 6 7 x =2BT-* ' Figure 18. Range Response For K = 0 Cube-Law Nonlinearity t 1-0....... 0.8-50%- Cubcr-.. C14 1--0,6-25% Cube 0.41 10% Cube 0% Cube 0.2 100% Trianglte 0 2 47 /! X =2B, 10 Figure 19. Range Response for K = 1 Cube-Law Nonlinearity
-10- _ oi -. O. 8 502 Cube--- 0J.4} 002." Cube 0% Cube I." - ' I 0.2 IOOZ Triangle Figure 20. Range Response for K = 2 Cube-Law Nonlinearity 1.0" t 0.8 50% Cube U 4-0.6-25 Cube 0.44 0 1 2 3 5 6 7 8 9 x= 2 BT- 0.2- Cube; 100Z Triangle 102 Cube,~~0 ub 0 1 2 3 4 57 8 9 X = 2BT + Figure 21. Range Response for K = 3 Cube-Law Nonlinearity 31
c~*1 10002 TCiagbe 0.- 0.2 2 10 Figure 22. Range Response for K~ =4 Cube-Law Nonlinearity x----- S 32
p % /1 0 Km C-A '0 0 004 LrI PI 4) ca n. bll* -4* OCCo z 33z
C 4 *- Cd i 0 -- A 0 C4U * -44 0)c -14 41 * u~ 4 -- 000 -'.4 344
5. EIPErU AL AD THEORErica VERIFICATION OF RESLLTS 5.1 Experimental Verification For the case of triangle plus sinusoidal modulation, the two limit results (i.e. pure triangle or pure sinusoidal modulation) are well-known. Since the theoretical range laws given in Figures 3-9 fall within these known limits, these results are already verified experimentally. For the case of triangle plus cube-law modulation, the theoretical range laws have been given by Figures 18-24. These results will now be verfied experimertally by using the test apparatus shown in Figure 25 below. The FM oscillator was supplied by HDL and is capable of large peak-to-peak deviations with little incidental AN. B was set to 20?Wz and fm was 5KHz. The experimental results are shown in Figures 26 and 27. Wide Band Oscillator 30021 RF Mixer Reflectot FM Output 300Q Transmission Line Mod. 5KHz k--h Harmonic -4150 Band Pass Envelope Strip Filter (Tuned Detector Chart to MHz) i IRecorder I I ExactCube-Law IExact m t J Function - Generator (5KHz) m Output I I KI m(t) Composite Output I --- ) Figure 25. Test Apparatus For Experimental Measurement of Range Laws 35
tic= Of0 Triangles Lim)[= 10O~=10%Cub La 0P = 24 ub a 1. 0 0 t = 100 Triangle Fiur 26XxeietlRng asfrk4- Tric=10ub angl Plus Postiv Cube-a Nolnarit T36_ Ot
K= 8 K=8 1OOPc = O+52 IOOP c 10Z+5Z Cu6e Law I." VL6 1OOPc ' 5% Cube Law loopc 50% Figure 27. Experimental Range Laws for K = 8-- Triangle Plus Positive Cube-Law Nonlinearity 37
It is seen that these experimental results agree with Figures 22 and 23 wher. the attenuation characteristic of the transmission line in the experimental apparatus is taken into account. 5.2 Theoretical Verification If negative percentages of cube-law modulation are used we have a modulating waveform of the form m() = k 1 t - k 2 t 3, -T/4 S t - T/4 (36) where the t = 0 axis has been shifted by T/4 for analytical simplification. This may be compared with the waveform shape that is obtained for triangle plus sinusoidal nonlinearity m(t) = k 3 t - k 4 t 3 + higher order terms (37) where the expansion sin 0 = e - (1/3!)03 + (l/5!)0s... has been used (see Appendix D). Neglecting the terms of order higher than t 3, a relationship between the percentage of cube-law nonlinearity and the percentage of sinusoidal nonlinearity can be obtained such that the two waveforms of (36) and (37) are equivalent; that is k, = k 2 and k 2 = k 4. Referring to Appendix D, the result is PC = 1 6(2/w) 3 P -0.64596Ps (38) Ps PS 6 (2/r) 3 13.304 6 (2/r) 2 (1-2/iT)-l which may be approximated by [PC = -0.646Ps] (39) when IPs1 < 1. Equation (39) is plotted in Figure 28. Using (38), some range laws for cube-law nonlinearity which are equivalent to the sinusoidal nonlinearity are shown in Figure 29. It is seen that Figure 29 is almost identical to Figure 4. Thus, the theoretical results for the cases of cube-law and sinusoidal nonlinearity are verified theoretically with respect to each other. 38
- I M IF TiT: I N.4. * - j~~ij* -~ ;Fk. 0. ) %D 04-1 _ > *- Ito i /r. u.: i - - I. 0 T-T P4' I 0 L114 + _ Ii 60 39
0.6-- 0.4-6.5 Cue( 00Aie 0.2 - t- _ 2' 3s 4I 01 1OO%~~ TiiBTI~ Fiur 29IoprsnofRneLw- I -La Nolnaiy 40I
6. SU~NaaM -Range las have been obtained for a class of distance measuring systems. This class radiates an RF signal which is frequency modulated by a nonlinear triangalar waveform. The nonlinezrity consists of one of two shapes-type I and Type I-as shown in Figure 30. Type I nonlinearity may be synthesized from a triangle plus a positive sinjsoidal nonlinearity or from a triangle plus a negative cube-law nonlinearity. Type II may be synthesized from a triangle plus a negative sinusoidal nonlinearity or a triangle plus a positive cube-law nonlinearity. (a) Type I Nonlinearity - i u(t) T/2 t (b) Type II Nonlinearity Figure 30. Nonlinear Frequency Modulating Waveforms The amount of nonlinearity is specified by a parameter which gives the * percentage contribution of the sinusoidal (loops) or cube-law (10OPc) nonlinearity with respect to the composite waveform amplitude at the time when the These percentages may have posi- peak value occurs (i.e. t = 0, T/2, T, etc.). tive or negative values. The resulting range laws show that the nonlinearity may be desirable or undesirable. The Type I nonlinearity produces an undesirable effect (Figures 3-9) since the sidelobe levels are increased over those obtained for pure tri- - angular modulation, for all distances greater than x 0. The Type II nonlinearity produces a desirable effect (figures 11-16 and 18-24) provided that the percentage of nonlinearity is not too large. That is, the sidelobe levels are -decreased for distances greatcr than xo if Ps or Pc is small and of the correct 4
I sign. From the figures it is seen that a sinusoidal nonlinearity with Ps = -1.5 (1/k+4) (40) o - - a cube-law nonlinearity ith Pc z l c+4) (41) gives the optimim sidelobe suppression and that the x 0 is not shifted very much for these small percentages of nonlinearity. In summary, the results of this report may be used to evaluate existing distance measuring systems and to design new systems. The range laws of existing systems may be predicted for the case of frequency modulation by a nonlinear triangular waveform. For new systems, modulating nonlinear triangular waveforms may be prescribed which produce range laws with better sidelobe suppression than those obtained by use of the pure triangular waveform. 42
APPENDIX A Fourier Series for the Mixer Output Equation (9) will now be derived from (3), (7) and (8). output is The mixer y(t,t) = Re { ck e j k w mt ej~ct }. (A-1) Expanding T(t) in a Taylor's series about t we have dti. A2 T(t 0 + t) = T(t 0 ) + d t 0 +.(A-2) Now dt AT (2vr/As) 2 vr 4 dtlt o At (c/as) c *1 where and AT is the incremental change in the round trip time delay during the incremental time At As is the incremental change in distance vr is the radial velocity of the target with respect to the DMS 2 c is the speed of light. From (A-2), using the first two terms of the Taylor's series Wct =0 + wdt (A-3) where o = CT(to) and wd = ( 2 vr/c)wc wd is the radian doppler frequency. angle 00, (A-1) becomes Using (A-3), but neglecting the phase y(t,t) = Re {Ck e (k ( om + wd)t} k = - = Re [coejwdt] + Re {[cke.ik"hmt + c-ke - j k("mt] e j ' )dt } (A-4) k=l 43~
We will now show that ck = c.k for the case when the instantaneous difference frequency has odd symetry; that is when wid(t,.t) = -wid(-t,' ) (A-5) From (8) Ck ft ejud(t,t ) ejkwmt dt (A-6) T 0 and 1 T ej~d(tt) jkwmt c-k = f e dt (A-7) Letting a = -(t + T), (A-7) becomes C-k T 0T ej'd[(t - o) ' T] e - jki m do (A-8) Identifying (A-8) with (A-6) we see that a sufficient requirement for c-k = c k is that Od[t,T] = Od[(T-t),T] and since 0 d is periodic with period T, this is equivalent to the requirement OdIt,T] = Od[-t,T] (A-9) Taking the derivative of both sides of (A-9) d(t,t) = do(t,t) _ -d~d(x,r) dt dx =x=-t Thus c k = c- k if id(t,t) = -wid(-t,t) (A-10) Using (A-10) in (A-4) we have 01 y(t,t) = Icol + i 21ckl cos [wdt + cos (kwmt) (A-I) k=l provided that wid(t,t) = -wd(-t,t) where I[']I denotes the magnitude of [' and ] denotes the angle of [. I4
Now let t o be defined as the value of t which gives the first turnaround of the modulation, m(t). This is illustrated in the figure below for the case of m(t) having a triangular waveshape. wid(t,t) to V t Then, using the above sufficient condition for ck = c- k, we have the requirement, from (A-5), that (id[(t-to), T] = -wid[-(t-t), T] (A-12) and (A-Il) becomes y(t,t) = Icol + 7 21ckI cos [wd(t-to) + ck] cos [kwa(t-t 0 )] (A13) k=l Furthermore, since wid is periodic with period T, (A-12) is satisfied if we require the difference frequency to have half-wave odd symmetry, that is if id(t,t) = - id(t + 2 T)-! A 45
APPEMDIX B Fourier Coefficients For Triangle Plus Sinusoidal Nodulation Using (6), (16) and (17), the difference phase is S(VaT)t + 2 -- [sin %(t-t) + sin wmt - sin wmt],0 S t S Od(t,'T) (-DaT)(t-T) + Db [sin WM(t-T) + sin wm*-sin wmt], < t < T (B-l) where the exact representation for 0 d during the turn-around time has been neglected. Referring to (18), the peak-to-peak frequency deviation in Hz is B = [2D(b + at(b-2) Using (18) and (B-2), (B-l) becomes (--)t + -- [sin wm(t-t) + sin WmT-Sin wmt], 0 < < t T T WM2 ed(t,r) = Bw : ( --)(t-t) + P [sin Wm(t-T)+sin, 1 -T-sinmt] < t S T T W2 (B-3) where and A Pt = 1 - Ps loops is the percentage of sinusoidal modulation loops is the percentage of triangular modulation. Using (B-3) in (11) we have for the Fourier coefficients 4PtT Ps T/2 JB)T (T)t + [sin wm Wm(t-T) + sin Wm - sin wmt Tk= 0 f e e- j kwmt dt _(4PiT Ps T jb )t + jm m [sin wm(t-t)+sin WMT - sin wmt]j + f e T,Jwtdt T T/2 (B-4) 46
47 1 Using a change of variable, t 1 = (t - the approximation: T/2), on the second integral and using [sin wm(t-t) - sin wt' z (wt) sin (w=t -T) r<< T (B-4) becomes Ck : 1 ejbwp s sin %+o ejbi (- --)t + PS T sin (wmt - e - k1 m t dt +(- 1 )k ejb[2pt+ T/2m e-jbiln - 2)q e - j k nt dt 0 (B-5) Using a change of variable, 0 sin aut =WmT for T << T, (B-5) reduces to = (2,/T)t, and the approximation k e Y[(! Psx),(k-Ptx)] + (-l)kejlptx Y[(- 2 Psx).(k + P ( (B-6) where Y(z~v) T e-j[z cos 0 + VOl 11') do (B-7) and A 2BT (B-8) Y(z,v) becomes Jv(-z) when v is an integer.
APPENDIX C Fourier Coefficients For Triangle Plus Cube-Law Modulation Using (6), (22), (28) and (29), the difference phase is 1(]CT)t + (Dlr)[ (t + vj T 1 0 t 6d(t,T) 3 3 T 3 ' (-DcT)(t-T) + (-DdT)[(t - T) - ()], S T (C-1) where the exact representation of Od during the turn-around time has been neglected and the following approximations have been used T { )]3 (t-t T (t 2 T T [t- (C-2) + ( ) T< 4 and 3 )] 3 33 [t -(3_ T + T)] 3 (t - T) - 3(t - T) 3 t, T << T (C-3) 4 4 4 The approximations are valid for our problem of interest i.e. when T << T. Substituting (C-1) into (11), the Fourier coefficients are 12/2j{(c~r~ +) + (4) 3 ]} T/2 j{ (C ) -DdT)[(t - T T) 3 -( )] - c k fe + w t T -J{((c-t)(t-T) + (DdT)L(t -.1T)3 ( T 3]) m T/2 (C-4) Using the change in variable, ti = T-t, on the second integral, (C-4) becomes T3 3 T/2 j{(dc)t + (DdT)[(t - )3 + (1)3]} Ck f e cos (kwmt)dt. (C-5) 48
Using the change of variable 0 - (2v/T)t, this reduces to -ICTT-.w [( $ej((=j- " ) 8 + (DIr)'73( -13+ (21)3]) ck 0+ (D))[ 2 2 cos(ke)do (C-6) ) Referring to (30), the peak-to-peak frequency deviation in Hz is TI B - f2d{[d 3 + c()] (C-7) Using (30) and (C-7), (C-6) becomes 1 Of e cos(ko)de (C-8) gk=o when x =2BT A (C-9) Pt= 1 - PC(C-) loopc is the percentage of cube-law modulation, loop t is the percentage of triangular modulation, x is the normalized round-trip time delay between the fuze and the target, and B is the peak-to-peak frequency deviation (Hz). 49
APPENDIX D Percentage of Modulation Required For Equivalent Sinusoidal and Cubs-Law Nonlinearity For triangle plus sinusoidal nonlinearity, using (16) and (17), modulating vaveform is IT u(t) =at +b sin wat It it - (D-1) the wbere the t - 0 reference has been shifted by T/4 for convenience. Then approximating the sinusoidal nonlinearity of (D-l) by the first two terms of the Taylor series we have a(t) = (a + bwm)t -t b(wt) t (D-2) 6 4 Identifying (D-2) with (37) we have and k 3 = (a + bwm) (D-3) k4= b (D-4) For triangle plus cube-law nonlinearity, using (27), (28) and (29), the modulating waveform is m(t) = ct + dt 3 t < (D-5) where the t = 0 reference has been shifted by T/4. Identifying (D-5) with (36) we have ki = c (D-6) k 2 -d (D-7) It is required that k I = k 3 and k 2 = k 4 for these two types of nonlinear modulation to be equivalent. Thus, (D-3) through (D-7) becomes c = a + bwm = a + 2b(D-8) 6 - b ((9) ( ) 50
Sojlvigg (US)Q, (D-9). (18) and (19) simultaneously for PC In term of Ps we have PCi CZ(j) Ps (D-6) where 1OO?. is the percentage of cube-law nonlinearity and loops Is the percentage of sinusoidal nonlinearity. Equation (D-6) can be arranged into a more useful form -[l/6(2ar)] 3 s -0.64596 Ps I - (2i0 1 -(PsIl3.304) 6(2/1)2 (l-2/w)-l which can be approximated by Pc -0.646Ps IPsI < 1. (D-8) REFERENCES: 1. Heistrom, C.W., Statistical Theory of Signal Detection, Pergamon Press, 1968. 2. Skolnick, M., Editor, Radar Handbook, McGraw-Hill, 1970. 3. Rowe, H.E., Signals and Noise in Communication Systems, Van Nostrand, 1965. 4. Rihaczek, A.W., Principles of High Resolution Radar, McGraw-Hill, 1969. 5. Peperone, S.J., "Analysis of the C-Class Fuzing Systems," Harry Diamond Laortoies Technical Revort No. TR-914, 1961. 6. Tozzi, L.M., "Resolution in Frequency-Modulated Radar4' Ph.D. Dissertation, University of Maryland, 1972. 51