2‘:
`3;
`
`E
`
`
`
`
`
`»m.m,_.;u-M.Nflfi»-‘J-uv-DrMJ\V',‘>k«",I~!;hQt§fl«:8\W<-K\£-F-¥‘$91P3!4"QN>l'I!fl'5M)h&«‘Lunh£*‘l’@lxdb‘-RVmxww-.m»wu.r«rIV."/Vm.swv...
`
`
`
`
`
`IEEE PRESS I
`
`'
`
`1976 Editorial Board»
`
`Allan C. Schell, Chairman «
`
`E’
`
`I
`
`0 George Abraham "
`Robert Adler
`Clarence J. Baldwin
`Walter Beam
`Mark K. Enns
`
`E. E. Grazda
`Robert C. Hansen
`
`R. K- Hellmann
`Edward W. Herold
`
`William G.7l'-Elowardf”
`Thomas Kailath
`Dietrioh’ Marcuse ..
`SanjitMitrar-x-
`James H.F.’omerem-:~I.
`
`_
`
`I‘
`
`_
`
`lrvin‘gI:Reingo|d-'I;v.+.’A
`Julian Reitmanuw
`
`A. ELI Siegmamr
`John‘B'. Singleton‘ :
`
`..
`
`v
`
`M
`
`’ W. P. Crone, Managing Edito‘rn<7"'I
`I
`I
`I if Carolyne_-Elenowitz, Supervisor, Publicati'r}n;~'Productionex: 9.
`
`'
`
`"
`
`Copyright © 1976 by
`THE INSTITUTE OF ELECTRICAL AND ELECTRONICS ENGINEERS-,»IlNCl,‘»
`345 East 47_Street, New York,;NY.,'.lOOl7"
`._
`. «
`'
`All rights reserved;
`V’
`'
`International Standard Book Numbers: rClothboLIrnd:30‘-879%Zi0'77;4ZI»"£j‘
`I, Pa perbou nd: Oje87-942-O7.8'iA2[.e
`
`Library of Congress Catalog Card Number 76"-18432
`
`PRINTED IN THE UNITED STATES OF AMERICA
`
`4'.v
`
`PETITIONERS 1014-0002
`
`

`
`A SPREAD SPECTRUM RANGING TECHNIQUE
`FDR AEROSPACE VEHICLES
`
`R. C. Dixon
`Member of Technical Staff
`TRW Systems Group
`Redondo Beach, California
`
`Abstract
`
`This paper considers the factors affecting
`spread spectrum ranging and describes a non-
`ambiguous, fast operating "hybrid" system util-
`izing a combination of high rate pseudonoise
`modulation for fine range resolution and low
`rate modulation for coarse range resolution.
`This technique allows rapid distance measure-
`ment while maintaining the inherent accuracy
`of other pseudonoise ranging systems.
`The
`pseudonoise code sequences used are much
`‘shorter than those conventionally required for
`nonambiguous long distance ranging, which faci-
`litates fast code phase acquisition. The,low
`rate "range message” used to resolve any ambi-
`guity due to repetition of the short code se-
`quence is derived from the repetition of the
`pseudonoise ranging sequence itself.
`
`Introduction
`
`Spread spectrum ranging systems offer an
`"excellent means of long range distance measure-
`ment.
`The capabilities of these systems have
`been spectacularly demonstrated in space probes,
`supplying information required for course,correc-
`tion and operational timing with great accuracy.
`The inherent long range accuracy of spread spec-
`trum ranging also makes it extremely attractive
`for navigational use in aircraft and manned space
`vehicles, but the relative size and complexity of
`first generation PN equipments has discouraged
`their use. Further, from an operational stand-
`point, spread spectrum systems have been imprac-
`tical to use under rapidly changing conditions due
`to the fact that most practical systems require
`synchronous code sequence operation (that is,
`the
`signal seen by the receiver must be precisely
`correlated in time with a locally generated
`reference signal) and the time required to
`achieve the desired synchronism is often such
`that a high speed aircraft or other aerospace‘
`vehicle can obsolete a range reading before it
`is'completed.
`
`Because of the unique circumstances under
`which deep space systems are used, relatively
`long code sequence phase acquisition times are
`allowable. That is, the information required,
`while possibly critical to mission success, may
`be supplied by a lengthy,
`time-consuming process,
`without Jeopardy to the mission itself, and syn-
`chronization times of up to several minutes have
`been common in space probe ranging systems.
`Therefore, deep space ranging techniques have not
`been directly applicable to highly manueverable
`vehicles. For aircraft use, where spread spectrum
`ranging offers its greatest utility, the implemen-
`tation methods used for spacecraft ranging are not
`feasible, as a range measurement must be available
`whenever it is desired and with minimum waiting
`time.
`
`trum equipments capable of being carried by almost
`any aerospace vehicle. Digital circuits, such as
`shift registers, often make up major parts of
`spread spectrum equipments. High speed shift re-
`gister implementation, for instance, has been im-
`proved over the years from the use,
`in 1959, of
`large lumped-constant delay networks to the pre-
`sent use of integrated logic circuits.
`A special
`integrated circuit capable of operation at bit
`rates in excess of 50 Mbps has been developed es-
`pecially for code sequence generation in spread
`spectrum systems. This unit includes in one pack-
`age both a flip-flop and a gated modulo-two adder,
`thus enabling construction of a shift register code
`sequence generator using only one package per stage,
`including all necessary feedback logic.
`In addition,
`for a given shift register length, every possible
`feedback combination is selectable. This makes
`' selective range addressing through code assignment
`quite easy to implement.
`
`Effort has also been expended in the area of
`building entire shift register sequence generators
`and other digital subsystems in a single integrated
`circuit package,
`through MOS techn1ques.- -~~
`
`Though PN ranging systems have not been reduced
`in complexity, microminiaturization techniques have
`improved their reliability and reduced size to per-
`mit their use invboth aircraft and space vehicles.
`Thus,
`time required for achieving synchronism bet-
`ween transmitting and receiving units has become
`the major factor limiting usage of pseudonoise
`ranging systems. Reduction of measurement time
`for ranging systems is the problem to which the
`hybrid ranging system described here is addressed.
`
`Factors Affecting Spread Spectrum Ranging
`
`_
`
`,Pseudonoise Ranging. Several types of spread
`,
`spectrum systems are in existence, each of which has
`its own particular advantages. All are characterized
`by the use of a much wider band of frequencies than is
`necessary for transmitting the basic information
`sent.
`Systems which hop from one frequency to
`another, or send short bursts of information at
`high rates, or simply modulate a carrier with a
`coded message, all are classed as spread spectrum
`systems.
`The type which has been used most often
`for ranging, however, is the direct sequence or
`pseudonoise (PN) modulated system. This is due to
`the relative simplicity of implementing PN systems
`(compared to other spread spectrum techniques) and
`to their_inherent synchronous operation. For these
`reasons, only PN systems are considered in the follow-
`ing discussions, even though the basic ranging tech-
`nique is applicable to other systems.’
`
`PN (pseudonoise) communications and navigations
`systems are characterized by their use of a pseudo-
`rendom code sequence to directly modulate an RF
`carrier. The codes used in PN systems generally
`have bit rates which are much higher than the bit
`One objection to spread spectrum systems, that
`of size, has been resolved by the introduction of
`rate of any information transmitted.
`(Typical bit
`rates for PN code sequences are in.the one to ten.
`integrated circuits. Both linear and_digital.inte-
`*7Mbps region). Signals modulated by’sudh'sequencés’
`grated circuits are applicable to spfieaa spectrum
`therefore are wideband and fit the general cate-
`systems, but the advent of high speed digital inte-
`gory of spread spectrum signals.
`grated circuits has made possible small spread spec-
`Reprinted from a paper in the 20th Annual Southwestern /EEE Conf. and Exhibition Flec., Apr. 17-19, 1968 and published by TRW.
`
`2,7 7
`
`PETITIONERS 1014-0003
`
`

`
`number of bits to be searched is large, and search
`time becomes prohibitive at long range.
`
`An alternative ‘.to' the transmiti-then-receive" 7
`(interrogate-respond) technique uses two frequen_
`cies,
`in a duplex arrangement such as that shown
`in Figure lb. Using this technique, only a fre-
`quency translation is necessary in the unit being
`ranged on. This unit retransmits the signal
`received (instead of a locally generated signal
`as in the responder) so that the unit measuring
`range receives a coded signal delayed by a number
`"of code bits equivalent to twice the distance
`being measured.
`The ranging system must contain
`two code sequence generators-one for transmission,
`and the other for matching to the received signal.
`Once the received signal has been matched to the
`receive sequence generator, range measurements are
`made by simply counting the number of bits delay
`between the transmit and receive sequence genera-
`tors in the ranging unit. Also, measurements may
`be continuous once system synchronization and an
`initial measurement occur. This is not true in
`the interrogate-respond system, as an interroga-
`tion and a response is necessary for each measure-
`ment. Again,
`the code sequence used must be longer
`than the number of bits delay at the longest range
`to be measured.
`»
`~
`
`Chief drawbacks to the duplex measurement
`method are that two sequence and clock generators
`are required in the ranging unit and the frequency
`allocation problem is worsened because two channels.
`are needed. For satellite applications, however,
`it is best because of the possible simplicity of
`the vehicle unit.
`,The hybrid system developed is‘
`applicable to both interrogate-respond and duplex
`systems.
`In both systems,
`the code offset which
`must be searched out for synchronization is equal
`to twice the actual range. This is an important
`consideration, for ituincreases resolution, but
`also_increases search time.
`
`The bit, or "clock,'»
`Code Sequence Bit Rates.
`rate at which the PN code sequence is generated
`determines the basic resolution capability of a PN
`ranging system.
`In'fact, resolution capability is
`limited only by the bandwidth required to support
`higher and higher bit rates.
`-As bit rate is
`increased, range uncertainty is decreased, but at
`the cost of a wider band RF channel. High resolu-
`tion ranging can be maintained, however, without
`x
`transmitting the entire (sin x)z spectrum, and
`many PN ranging systems transmit only the main lobe,
`or less, even though the biphase modulated RF signal]
`envelope may be seriously degraded by such narrow-ban
`ing.
`
`
`
`The fact that basic resolution capability is
`not degraded by loss of bandwidth is due to the
`correlation properties of the code sequences.
`q
`Synchronization between received and locally gen-
`erated code sequences can be judged on the basis of’:
`correlation over the whole sequence length, rather
`than on a bit-by-bit basis. Therefore, a degrada-
`tion in rise and fall time does not affect measure-
`ment accuracy, as long as the bandwidth.is enough
`to reasonably reproduce the transmission of sinzlev
`code bitS- Bandwidth loss effects on hybrid system
`are identical to other PN systems.
`
`PN code sequences have a correlation function
`which ideally is an
`isosceles
`triangle in which
`the base width is
`equal to twice the width of
`a single code bit. Exact code synchronization
`corresponds to the peak of this correlation func‘
`tion, while an advance or delay between received
`
`2713
`
`PETITIONERS 1014-0004
`
`PN codes vary in length from a few bits to
`billions of bits, depending on their use.
`The
`pattern of ones and zeros in such sequences doea_
`*f‘not repeat within this code length, and thus to .’I'
`an observer who views less than the entire se-
`quence,
`the appearance is that of a random set
`of bits, such as might be produced by a noise
`generator. This noiselike stream of bits is
`actually a well defined signal whose short
`term properties resemble noise but which is
`repetitious and therefore quite predictable.
`(Hence the name "pseudo" noise). Signals such
`as these are simple to generate with shift re-
`gister sequence generators, and these are used
`in almost all PN systems to construct the basic
`spectrum-spreading signal.
`The specific code
`sequence used, its,length, and bit rate must
`"be decided by bverall system constraints of
`accuracy, bandwidth, and measurement time.
`
`.
`
`Figure 1 illustrates a typical PN ranging
`system.
`In the transmitter, a carrier is biphase
`modulated by the code sequence generator, which
`produces a wideband, suppressed carrier output
`signal, code modulated.
`The modulation spectrum
`produced by a PN code modulated transmitter is
`(sin x)2’ with a main lobe bandwidth (null-to-
`x
`"
`'
`"
`'
`'
`"
`
`null) equal to twice the code sequence's clock
`rate.
`
`The receiver "decodes" a received signal
`by matching an internally generated sequence to
`the incoming signal. For ranging,
`the important
`point is that the receiver matches its internal
`sequence to the signal that it sees, and this
`signal is delayed by an amount of time equal to
`the propagation time from transmitter to receiver.
`The transmitter,
`then, is generating the same code
`sequence as the receiver, but is a number of code
`bits ahead of the receiver. This number of code
`bits depends on both distance and the rate of
`code generation.
`- ~~-
`
`”
`
`Once the transmitting system has transmitted
`for a period long enough to allow the receiver to
`match its internal reference to the signal trans-
`mitted,
`the transmitter can send information, or
`instruct the receiver to transmit back to the
`original transmitter (called "interrogation" and
`"response"). Now, if the interrogating unit
`transmits a mark to the responder which causes
`the responder to switch to the transmit mode,
`and the interrogator goes into receive,
`the two
`are no longer synchronized.
`(Remember that the
`original receiver was "behind" the transmitter).
`The interrogator must now delay its code sequence
`so that it is "behind" the responder by the number
`of code bits equal to the propagation time. All
`that is now necessary to allow the interrogator
`to measure range is to count the number of bits
`that it must delay its code sequence to bring it
`back into synchronization with the unit now trans-
`mitting.
`(Note that the range measurement is made
`as a function of counting bits, and is a digital,
`discrete measurement accurate to within one bit
`period. Therefore,
`the accuracy of measurement
`is the same at any range, where the threshold of
`the system is exceeded).
`
`Systems of this type have been implemented,
`and have been shown to be accurate and reliable.
`The limiting factor in the use of such a simple
`system is that the code sequences used must be
`long enough that they do not repeat over the
`maximum distance measured.
`(Otherwise, an ambi-
`guity exists which is not resolvable from simple
`code sequence offset measurements). Thus,
`the
`
`

`
`the interrogate-respond and duplex ranging systems,
`the number of code bits of offset due to range is
`actually equivalent to twice the range being mea-
`sured. Therefore,
`in these and similar systems,
`clock rate required to give a desired resolution
`is halved.
`
`Once the nominal clock rate is chosen, both
`transmit and receive systems are expected to
`operate at the chosen frequency (within their
`error limits) except when a receiving unit is
`-in the "search? mode,
`looking for correlation
`with an incoming signal.
`
`Searching is accomplished by operating the
`receive unit clock at a rate slightly higher or
`lower than the nominal clock rate. This causes
`the receiver's code sequence to either advance
`or be retarded in phase with respect to the trans-
`mitted sequence,
`thus "sliding" the two sequences
`past one another comparing all possible code phase
`relationships until correlation occurs.
`In the
`interrogate-respond ranging system discussed pre-
`viously, this search procedure occurs twice:
`first, when the initial transmission occurs, and
`second, when the responding unit replies.
`
`-Direction of search (forward or backward in
`code phase) is usually not important to a PN sys-
`tem, but in transmit-receive ranging systems it
`is important that the unit making the range mea-
`surement (the interrogator) search backward. That
`is, the interrogator's clock should run at a rate
`lower than the responder's clock, so that its code
`sequence is slipping backward with respect to the
`incoming signal. This is important,
`in that the
`responderls sequence generator is retarded with
`respect to the interrogatorfs sequence generator
`at the time of the transmit to receive switch,
`and the interrogator's code sequence must also
`be retarded to a number of bits equal to the
`propagation time in order to recorrelate.
`The
`period during which the interrogating unit is
`searching for recorrelation with the responding
`transmitter is the most vulnerable for error
`entry. During this time, any relative clock
`drift or doppler shift is accumulated by the bit
`search counter in the interrogator as a range
`error. This error is a function of relative
`clock difference, relative velocity, Search time,
`and the difference between desired and actual
`clock rate,
`(Where actual operating rate is
`determined by crystal oscillators the clock can
`be held within a few cycles of the desired rate,
`so that errors due to systems operating at other
`than the desired nominal rate are negligible).
`On the basis of clocks operating near their
`design frequency, range error for an integral
`frequency clock system may be expressed by:
`
`_
`(iIF:r:::s)— (
`where:
`
`_
`-3
`[KI KR] 431.7 VR KR x lo ) TS
`
`KI= receive unit clock rate (Bps)
`KR= transmit
`clock rate (Bps)
`vR= relative velocity (N.M.P.H.‘)
`T5:
`correlation search time (seconds)
`
`This expresses the errors accumulated due to a
`receive unit counting at an assumed clock rate K ,
`while transmitter's reply actually is at a clock
`rate KR offset by the doppler frequency shift
`referred to transmitter clock signal. This is
`multiplied by‘search time, for the_longer thef
`time of the receive unitls search for correla-
`tion,
`the longer is the time that errors are
`
`.279
`
`PETITIONERS 1014-0005
`
`3
`
`;
`e
`9
`
`and reference sequences decreases correlation,
`which in effect is similar to raising the re-
`ceiver threshold.
`The important point is that
`received and locally generated references must
`be within one bit before recognition of syn-
`chronization is possible.
`In fact, it is
`necessary in most systems for code synchroni-
`zation to be within a small fraction of a bit
`of the maximum point of correlation before the
`code correlation circuit's output signal—to-
`noise ratio is sufficient to permit recogni-
`tion of synchronization.
`For this reason,
`synchronous PN systems are inherently incap-
`able of measurements in error by more than
`one bit. Narrowing of transmitter or recei-
`ver bandwidth tends to round the edges of the
`triangular correlation function as in Figure 3b,
`but this effectively raises receiver threshold
`rather than broadening the correlation area.
`Again,
`inherent accuracy is affected negligibly.
`what,
`then, does affect PN system ranging accur-
`acy?
`
`Selection of Clock Rate. Factors affecting
`the accuracy of a PN ranging system are chiefly
`clock rate, clock drift, and doppler shift of
`the clock rate.
`(Ignoring possible errors in
`the digital counting and readout circuits,
`which should be negligible).
`
`Clock rates may be the same whether a
`ranging system is of the hybrid or conven-
`tional type.
`It is necessary to give some
`consideration to the combination of the clock
`rate in conjunction with the PN sequence length
`in a hybrid system, however, as the range message
`bit rate is the result of these parameters.
`
`A PN system using a code sequence bit rate
`vof 1.618750 MHZ would therefore have a propaga-
`tion delay of 10 bits per mile and a receiver
`10 miles away would see a signal delayed by
`exactly lOO bits from the transmitter. That is,
`the signal appearing at the receiver was trans-
`mitted at a time which, on the transmitter's
`terms, was 100 bits earlier. Therefore,
`the
`receiver's sequence generator must be exactly
`100 bits behind the transmitter to achieve maxi-
`mum correlation, and between 99 and 101 bits
`behind the transmitter to be within the bounds
`Of correlation.
`1.618750 MHZ is certainly not
`a standard frequency, but an accurate crystal
`stabilized source at this frequency is readily
`achievable.
`(As an aside,
`it" should be noted"
`‘that multiples of .l6l875 MHZ may-be used as
`clock rates to gain whatever resolution is '
`desired, i.e. one-fiftieth mile would require
`a clock frequency of SO x .l6l875 = 8.093750
`,Mbps. Of course,
`the number used for C,
`the
`Speed of light, is that for light traveling
`‘in a vacuum, so that in any other medium
`‘Some error will be seen. This is not signi-
`icant, however, as long as all clocks oper-
`ts at the same rate. Oiluper'cent range*w4,’
`ierror limit due to common frequency offset
`iflould allow the clock rate to change up to
`618 bps).
`It is recalled that,
`in both
`
`
`
`For example, let us assume that we wish
`to design a ranging unit capable of measuring
`range to within 0.1 nautical miles (one n.m. =
`1852 meters), and that we want to count 10 bits
`per mile of propagation delay. 0.1 mile corres-
`ponds to a bit rate having a wavelength of 185.2
`meters, or a repetition rate:
`8
`3:95; 25 :12 =
`speed of light
`resolution
`
`we
`
`IIII
`
`<D‘m‘-5m 0
`
`

`
`accumulated at this rate. Doppler frequency errors
`may reduce errors due to clock differences or add
`to them, depending on the direction of errors and
`relative velocity.
`
`Selecting the Ranging Code. Code sequences
`used in ranging should be chosen on the basis of
`considerations for their auto and cross correla-
`tion properties, and in most ranging systems, for
`sufficient length.
`
`A code sequence such as the maximal linear
`type is an excellent choice for ranging. Not
`only are the properties of these codes well known,
`but the methods for generating them are straight-
`forward and simple.
`'
`
`Where_codes other than maximal are to be used,
`.their auto-correlation and cross correlation proper-
`ties should be carefully analyzed. Otherwise, inter-
`ference between different sequences can nullify
`selective addressing, or minor correlation in a
`correct sequence can cause erroneous measurements
`due to falsely recognizing synchronization. Maximal
`codes are, by definition,
`the longest codes which
`can be generated by a given shift register, or a
`delay element of a given length.
`In the case of
`shift register sequence generators, which are the
`only type to be considered here,
`the maximum length
`sequence is 2“ -1 bits, where n is the number of
`stages in the shift register.
`A shift register
`' sequence generator consists of a shift register
`working in conjunction with appropriate logic,
`which feeds back a logical combination of the state
`of two or more of its stages to its input.
`The
`output of such a sequence generator, and the con-
`tents_of its n stages at any sample (clock) time,
`is a function of the outputs of the stages fed
`back at the previous sample time.
`
`Basic properties of maximal codes are:1’2
`
`The number of one's in a sequence equals
`l.
`the number of zero's to within one bit. For a
`1023 bit code,
`there are 512 one's‘and 5ll zero's.
`Therefore, any DC correlation term is negligible.
`
`The distribution of one's and zero's is
`2.
`well defined and always the same. Relative posi-
`”tions of runs of one's and zero's vary from code
`sequence to code sequence, but the number of each»
`run length does not. There are exactly Zn-(p+2)
`runs of length p, for both one's and zero's,
`in
`every code sequence.
`(Except for n-l zero's and n
`one's of which there is one run each. Also, n
`zero's and n-l one's of which there are no runs.)3
`
`3. Auto-correlation of a maximal linear
`code is such that for all values of phase shift,
`the correlation value is -1, except for the zero
`+ l bit phase shift area, where correlation varies
`linearly from the -1 value to 2“-l (the sequence
`length).
`A 1023 210-1) bit maximal code there-
`fore has a peak to average auto-correlation value
`of 102k, a range of 30.l db in signal.
`It must
`be realized, however, that these values for auto
`correlation are valid only for averaging over the
`entire sequence length.
`
`All of the properties of maximal sequences
`can be used to advantage in a ranging system.
`The question,
`then,
`in general, is not what type
`of code to use, but what length.
`
`A ranging code used by a system which has
`no secondary resolving capability must be of
`sufficient length that it does not repeat over
`the maximum distance measured. That is, to
`
`measure a range of 1000 miles with a code giving 50
`bits per mile resolution would require a code length
`of at least 50,000 bits. Otherwise, the code sequence
`would repeat, and the interrogator
`could recognize_
`synchronizationsat more than one range. At the .
`8.093750 Mbps rate required to give 50 bit per mile
`resglution,
`the repetition period of the 65,535
`(21 -1) bit code required would be 65 355
`= 8.1
`3-093750
`milliseconds and the bandwidth of the correlation
`detector would be #3 Hz.
`If a #3 Hz recognition
`bandwidth is used, however, searching out a range
`at the maximum distance could require almost 20
`minutes.~ Obviously,
`then in a simple ranging system,
`high resolution and long distance measurements must
`be traded off against measurement time.
`
`'
`
`
`
`The major problem in PN ranging for aircraft and
`maneuverable spacecraft has been the requirement for
`measurement at long range in a reasonably short
`period. Reduction of ranging time is limited by the
`maximum search rate a receiving unit is capable of
`achieving and the length of the code to be searched.
`Maximum search rate,
`in turn, is limited by the
`recognition time of the receiver's correlation
`detection circuits.
`(The receiver must be able
`to recognize correlation and stop the search pro-
`cess before the point of code synchronization is
`passed.)r To complete the circle,
`the bandwidth-of
`the correlation detectors must be commensurate with
`the auto correlation requirements of the codes used.
`
`What techniques are available to reduce rang-
`ing time while maintaining accuracy? Much work has
`been performed in the area of developing special
`"acquirable" codes which have the required length
`for long range measurement, but which also have
`synchronization properties such that a range may
`be searched out without traversing the entire code
`length.l
`Jet Propulsion Laboratories has had great
`success with code sequences made up of three com-
`ponent codes assembled in such a way that the over-
`all ranging sequence has a length which is the
`The
`product of the three component code.lengths.
`search for correlation is done in parallel, in three
`correlation detectors, each of which compares an
`incoming product—code-modulated signal with one of
`the component codes. That is, all phases of each
`component code are searched for correlation.
`
`in this process, is sequen-
`Synchronization,
`tial, with a partial correlation occurring each
`time one of the component codes reaches its point
`of synchronization and complete correlation occur-
`ring when all three are synchronized. This tech-
`nique can reduce range measurement time tremendously
`but suffers from threshold degradation due to par-
`tial correlation and attendant loss of correlator
`output signal-to-noise ratio when less than all of
`the component codes are synchronized.
`
`Hybrid Ranging System
`
`A ranging technique which solves the problem
`of fast range measurement at long distances has
`been developed. This is a hybrid system approach
`in that more than one kind of modulation is used
`to measure range.* In this technique, a PN code,
`a few thousand bits long, and a digital "range
`message" (whose bit rate is the repetition rate of
`the PN sequence), simultaneously modulate the trans-
`mitter.
`The PN code does not determine maximum
`range so that its length is chosen short, to reduce
`search time.
`(The only bound on reducing code length
`is that any system crosscorrelation requirements be
`met.) Ambiguity in range caused by using a
`short
`PN code is resolved by the low rate digital message
`whose length is such that its repetition period is
`
`*A similar technique was suggested by W. J. Judge of Magnavox Research Labs in an unpublished memo,
`
`in 1963.
`
`280
`
`PETITIONERS 1014-0006
`
`

`
` longer than the propagation delay of the longest
`
`No separate search process
`range to be measured.
`13 needed to synchronize this range message, how-
`ever, as it is clocked by the PN code repetition
`and is therefore synchronized by the initial
`search process. All that is needed is a measure-
`ment of the relative phases of local or received
`range messages, which is as simple as counting the
`number of PN code repetitions between range message
`markers. Figure 2 illustrates the relationship be-
`tween the PN code and the range message as they are
`used in the hybrid system. Each complete repetition
`of the high bit rate PN code sequence corresponds to
`one bit in the range message.
`The ratio of the bit
`rates is thus seen to be equal to the number of bits
`in the PN code.
`
`The range message itself is the simplest kind
`of digital signal, made up of a number of square
`waves whose half-wave period is the PN code repeti-
`tion rate, and which are inverted in phase after a
`time greater than the maximum expected propagation
`delay. This phase inversion is the marker which
`resolves the range ambiguity.
`If error correction
`capability is required,
`the range message can be
`structured to give the desired properties.
`The
`simple range message used was chosen to permit
`measurement without a separate message synchroni-
`zation process.
`
`Modulation is such that the PN code biphase
`modulates the transmitted carrier at its basic
`rate, permitting high resolution.
`The range
`message, however, is sent as a baseband signal,
`keying the same carrier that is modulated by the
`PN code. Composite transmitted output then is a
`fin x)2 spectrum whose carrier shifts at the range
`X
`_
`,.nm
`message rate. Either frequency or phase shift key-
`ing is satisfactory for use in range message modula-
`tion. Phase shift keying is most optimum, and may
`be implemented by inverting the PN code at each
`range message one - zero transition.
`PSK demodula-
`tion is much more complex than FSK demodulation,
`however, as a simple phase-lock loop or a dis-
`criminator may be used for FSK detection. Signal-
`to-noise ratio at the demodulator in either the
`FSK or PSK case is excellent, since the PN code
`used is fully correlated by the synchronization
`process. Therefore, either modulation technique,
`is satisfactory.
`
`Timing, and a range unit block diagram, for
`a duplex hybrid ranging system is shown in Figure
`3.
`A range measurement using this system would
`be made as follows:
`
`1. At the initiation of the measurement,
`both transmit and receive PN generators
`are reset and started.
`The transmit
`generator PN modulates the transmitter
`at the nominal PN rate, and the receive
`PN generator searches backward.
`
`2. As soon as the receive PN generator
`searches out the number of bits corres-
`ponding to range, synchronization occurs,
`and the number of bits searched.is entered
`into a range counter.
`.
`‘
`r
`
`the receiver sends a
`3. At synchronization,
`logic signal to the transmitter, which
`begins to generate the range message and
`to modulate the carrier with it.
`(Range
`message output from the receiver now
`delayed by the two-way propagation time
`..between the transmitter and receiver, and
`*~can"be'compared.with'the‘transmitted '
`message to resolve correct range).
`
`The number of PN sequence repetitions
`between the occurrence of a phase re-
`versal in the transmitted and received
`range message is counted, weighted, and
`added with the bits counted in the ori-
`ginal search process.
`The measurement
`is complete.
`
`Subsequent measurements of range and
`range rate may be made by counting the
`number of PN code bits between the
`occurrence of all ones vectors in the
`transmit and receive PN generators.
`
`Figure h shows simplified block diagrams and range
`message timing for a hybrid interrogate - respond
`ranging system which could be used where range
`readout is desired by both participants.
`It is
`recalled from Figure 2 that the entire PN sequence
`repeats once during each bit time in the range
`message, and is synchronous with it. Aircraft or
`manned space vehicles users might find the interro-
`gate - respond technique best because it divides
`system complexity equally between identical systems.
`
`One range cycle is described as follows (see
`timing diagram):
`
`1.
`
`The interrogating unit transmits a PN
`modulated signal for a period long
`enough to assure that the intended
`responder has synchronized.
`
`The range message, consisting of a
`_ series of one-zero hits (a square wave)
`at the PN code repetition rate,
`is sent,
`followed by a phase inversion of the
`range message. This information FSK
`modulates the PN modulated carrier.
`
`At the PN code repetition marker follow-
`. ing the phase inversion several things
`occur;
`(1) the responder starts genera-
`ting an identical range message to that
`sent by the interrogator (i.e. a given
`number of square wave periods, an inver-
`sion, repeating),
`(2) the interrogator
`switches to receive, and (3) the res-
`ponder switches to transmit.
`
`Search processing for resynchronization
`is delayed in the interrogator for a
`short period sufficient to exceed the
`T-R switching time.
`In t