.
`United States Patent
`
`[19]
`
`US005414729A
`[11] Patent Number:
`
`5,414,729
`
`Fenton
`
`[45] Date of Patent:
`
`May 9, 1995
`
`[54] PSEUDORANDOM NOISE RANGING
`RECEIVER WHICH COMPENSATES FOR
`MULTIPATH DISTORTION BY MAKING
`USE OF MULTIPLE CORRELATOR TIME
`DELAY SPACING
`
`FOREIGN PATENT DOCUMENTS
`
`7/1988 European Pat. Off.
`0312193
`0351156 7/1989 European Pat. Off.
`2215539
`3/1988 United Kingdom .
`
`.
`.
`
`[75]
`
`Inventor:
`
`Patrick Fenton, Calgary, Canada
`
`OTHER PUBLICATIONS
`
`[73] Assignee: NovAtel Communications Ltd.,
`Canada
`
`[63]
`
`[21] APPL N°-‘ 1581511
`[22] Filed:
`Nov‘ 29’ 1993
`_
`_
`Related U.S. Applicanon Data
`Continuation-in-part of Ser. No. 825,665, Jan. 24, 1992,
`abandoned.
`1.11. c1.6 ............................................. H04B 15/00
`[51]
`[52] U.S. C1. ......................................
`:33’/81's//241139;3342521)//33547;
`[ss] Field of Search ....................... 375/1; 330/34, 46;
`342/357
`
`l55l
`
`_
`References Clted
`U.S. PATENT DOCUMENTS
`
`1/1984 Gorski-Po iel
`4,426,712
`4,457,006 6/1984 Maine ‘
`P
`4,530,103
`7/1985 Mosley’ Jr. et a1_ _
`4,611,333 9/1935 Mccauister et a1_ _
`4,706,286 11/1987 Sturza .
`4,754,465 6/ 1988 Trimble .
`-
`:~73§:‘3fS3 1;/1933 $3‘; at 31-
`57?] 1
`Tac1tl1;‘t:”‘et'al.
`4,807,256 2/1989 Holmes et al.
`4,321,294 4/1989 Thomas, J1._ _
`.
`4,894,842
`1/1990 Broelchoven et al.
`5,101,416 3/1992 Fenton et al.
`........................... 375/1
`235
`
`.
`
`.
`.
`
`Langley, R- 3-, “Why is the GPS Signal 50 C0mP1eX7’’,
`GPS World, May/Jun. 1990, pp. 56-59.
`Article entitled Proceedings ofthe Satellite Division In ter-
`national Technical Meeting, Sep. 19-23, 1988.
`Primary Examiner—David C. Cain
`Attorney, Agent, or Firm—Cesari and McKenna
`
`Cr
`ABS
`[57]
`A receiver for pseudorandom noise (PRN) encoded
`3:193:33n§;j{;*;1§§xfi:i;;m§i§§1§t:Ir:f1g-;;j3};;g}§;§;§;:§
`Cime1:tors_ Tlhegampling_CifC}1it£T0Vid€Shdi%it:1 sam-
`P 35 ° 3 ,‘e°e‘Ve
`°°mP,°5“'F 51%“
`‘° C“ °, ‘. 8 56.“
`eral receiver channel circuits. The synchronizing cir-
`cuits are preferably non-coherent, in the sense that they
`track any phase shifts in the received signal and adjust
`the frequency and phase of a locally generated carrier
`reference signal accordingly, even in the presence of
`.
`.
`.
`.
`.
`Doppler or ionospheric distortion. The multiple corre-
`lators in each channel correlates the digital samples
`with locally generated PRN codes having multiple
`offsets, to produce a plurality of correlation signals. The
`plurality of correlation signals are fed to a parameter
`estimator, from which the delay and phase parameters
`0:1the diregt patth sitgnjal, acsl xgell as multipath sig-
`n 5’ may 6 es ‘ma 6 1 an
`°“‘ W 1°
`3 range ma‘
`surement may be corrected.
`
`19 Claims, 10 Drawing Sheets
`
`240-1
`
`ls
`ROTATE
`E CARRIER
`
`N9: 121 I K2
`
`
`
`EDOFP
`CODE
`W CARRIER AND
` SYNC (FIG. 3)
`
`
`
`
`
`S09'lOH.lNO3
`
`
`
`S08'l3NNVHO
`
`
`
`'
`
`,
`
`CORRELATOR
`1
`(FIG. 4)
`
`MULTIPATH
`PARAMETER
`ESTIMATOR
`
`oh. ‘lk
`
`
`
`
`PRN CODE
`
`n
`
`1:0
`1,: 27% . in, (xx .1,)cos(o,,)1o E5111,
`n
`
`(1,, .1, + 5.) cos (9., ..:,,,1)
`
`01:
`
`21’ “£0 ~ [R1 (IR 91') SIN (ok)]+ E a]R' (I3 9!. :8 |) SIN (ok 09"“)
`1:0
`
`PETITIONERS 1004-0001
`
`

`
`10
`
`ANTENNA
`
`V
`
`11
`
`Cs
`
`12
`
`IF
`
`DOWN -
`couvennsn
`
`20
`
`CHANNEL BUS
`
`.
`
`14
`
`20
`
`'3' as
`
`223
`
`CHANNEL
`1
`
`CHANNEL
`2
`
`CHANNEL
`N
`
`’
`
`18
`
`1:46
`
`CONTROL aus
`
`INTERFACE
`
`INTERHUPT
`
`
`
`1U91‘3¢I'S'f1
`
`g‘<
`:0
`
`H 8
`
`0'
`
`9
`31-9-
`l-I
`OH:
`5
`
`.,.
`-A
`D-l
`3:-
`x]
`
`N\
`
`o
`
`PETITIONERS 1004-0002
`
`166
`
`162
`
`nmen
`
`MEAS
`
`1667
`1-
`I
`I
`I
`'
`‘
`|
`'
`I
`|
`|'
`»
`n
`L______.____._______ _ _ _ _ .____.. _ _ _ # __.|
`
`DATA
`
`ADDR
`
`CTRL
`
`
`
`

`
`
`
`1I11913cI°S'f1
`
`=3
`xg
`
`E
`\g
`U1
`
`mD
`
`"to
`(D(‘F
`
`N OH
`
`:
`
`E
`
`U‘
`so
`.§
`3
`uq
`
`IQ
`\O
`
`PETITIONERS 1004-0003
`
`240-‘!
`
`CORRELATOR
`2 ’
`(FIG. 4) 270
`
`1
`
`_
`
`I1. 01
`
`MULTIPATH
`
`
`
`I
`
`"’
`
`O
`
`"
`
`ESTIMATOR
`
`on].
`
`1. 45M’.
`
`¢k- ‘k
`
`10. CD
`PRN :2
`
`connE|_A1on
`2
`
`, O
`
`I
`
`2
`
`2
`
`240-3
`
`ID! GD
`
`CORHELATOR
`3
`
`PEN Im
`
`COHHELATDH
`n
`
`n
`
`xi: /21% . gn, (zk ». 1,) cos (¢k)1 + :d'H'
`[:0
`
`(xk +1, + 5,) cos («pk +¢m1)
`
`GI: I21’ -‘Ts;
`
`. [H' (TR + 1') SIN
`
`n
`
`"
`
`+
`
`ulflt (‘Ck -4-“ +5 |) SIN
`
`CARRIER AND
`com:
`svnc (FIG. 3)
`
`PHN CODE
`
`,
`
`>
`
`OO 3:
`
`nOr
`
`- WCU
`
`:
`
`OI>Z2I
`
`11I''''
`
`WC(
`
`I)
`
`

`
`TO PRN com:
`GENERATOR
`
`
`nsw
`
`cu<
`
`‘
`
`'
`
`2-—-i\—-—"¥
`W0 “1"“P
`
`snow.
`CHANNEL aus
`F
`
`3
`
` INT4
`
`
`
`111913¢I'S°I1
`
` 01JO9was9661‘658w
`
`
`
`CONTROL BUS
`
`18
`
`I G)
`
`on
`
`6ZL‘17I17‘S
`
`PETITIONERS 1004-0004
`
`

`
`U.S. Patent
`
`May 9, 1995
`
`Sheet 4 of 10
`
`5,414,729
`
`PETITIONERS 1004-0005
`
`

`
`U.S. Patent
`
`May 9, 1995
`
`Sheet 5 of 10
`
`5,414,729
`
`DATA BIT
`
`T PRN CODE EPOCHS
`
`PRN CODE EPOCH
`
`Z PRN CODE CHIPS
`
`PRN cone CHIP
`
`N Fs CLOCKS
`
`I \\\
`1
`!
`
`\\\
`
`\\\\
`
`l\
`
`n\\
`2
`
`\
`
`1 I
`
`CARRIER WAVELENGTH
`
`PETITIONERS 1004-0006
`
`

`
`
`
`6zL‘17[17‘g01J09wellsS66!‘6KENmama-S-Q
`
`PETITIONERS 1004-0007
`
` D
`
`-0.4
`
`-0.2
`
`0.2
`
`0.4
`
`0.6
`
`OFFSET (CODE CHIPS)
`
`FIG. 6
`
`DIRECT PATH SIGNAL
`
`CORRELATION
`
`
`
`NORMALIZEDPOWER
`
`RESULTING SIGNAL
`CORRELATION
`DIRECT + MULTIPATH
`
`MULTIPATH SIGNAL
`CORRELATION
`
`PHASE = 0 DEG.
`DELAY = 0.2 CHIPS
`AMPLITUDE = 0.5
`
`

`
`
`
` 6ZL"l7I'l7‘S013°L19°'ISS66!‘6Knew1II91BJ'S'fl
`
`MULTIPATH SIGNAL
`CORRELATION
`
`PHASE = 180 DEG.
`DELAY = 0.2 DEG.
`AMPLITUDE = 0.5
`
`OFFSET (CODE CHIPS)
`
`FIG. 7
`
`PETITIONERS 1004-0008
`
`DIRECT PATH SIGNAL
`
`
`
`CORRELATION
`
`RESULTING SIGNAL
`CORRELATION
`DIRECT + MULTIPATH
`
`-
`
`
`
`NORMALIZEDPOWER
`
`

`
`
`
`6ZL‘17I17‘S01J081°°IISS661‘6mm1ua1ed'5'[1
`
`PETITIONERS 1004-0009
`
`DIRECT PATH SIGNAL
`CORRELATION
`
`RESULTING SIGNAL
`CORRELATION
`DIRECT + MULTIPATH
`
`
`
`MULTIPATH SIGNAL
`CORRELATION
`
`0 DEG.
`PHASE
`DELAY _ 0.2 CHIPS
`AMPLITUDE = 0.5
`
`-0.4
`
`-0.2
`
`0
`
`0.2
`
`0.4
`
`0.6
`
`0.8
`
`1.0
`
`OFFSET (CODE CHIPS)
`
`FIG. 8
`
`
`
`NORMALIZEDPOWER
`
`

`
`U.S. Patent
`
`May 9, 1995
`
`Sheet 9 of 10
`
`5,414,729
`
`moh<._m..._.ooo
`
`m._.<._
`
`moh<4mmmoo
`
`4<=»oz:m
`
`mg
`
`wz_v_o<m»
`
`mommme
`
`4
`
`mo»<._mmmoo
`
`>.E<m
`
`H3MOd CIEZITVWHON
`
`ad
`
`wetoo
`
`2..
`
`Nd.
`
`
`
`62:0mnoo.Emuio
`
`a.O_n_
`
`PETITIONERS 1004-0010
`
`
`
`

`
`U.S. Patent
`
`May 9, 1995
`
`Sheet 10 of 10
`
`5,414,729
`
`1.0
`
`0.8
`
`0.6
`
`0.4
`
`0.2
`
`
`
`CORRELATORS0-2
`
`0
`
`
`
`-0.6-0.4-0.2o OFFSET(coneCHIPS)
`
`-0.8
`
`!:l3MOd CIEZFIVWHON
`
`-1.0
`
`PETITIONERS 1004-0011
`
`

`
`1
`
`5,414,729
`
`PSEUDORANDOM NOISE RANGING RECEIVER
`WHICH COMPENSATES FOR MULTIPATH
`DISTORTION BY MAKING USE OF MULTIPLE
`CORRELATOR TIME DELAY SPACING
`
`CROSS-REFERENCE TO RELATED
`APPLICATION
`This application is a continuation—in-part of a prior
`U.S. patent application Ser. No. 07/825,665 entitled “A
`Pseudorandom Noise Ranging Receiver Which Com-
`pensates for Multipath Distortion by Dynamically Ad-
`justing the Time Delay Spacing Between Early and
`Late Correlators” filed Jan. 24, 1992, which is assigned
`to NovAtel Communications, Ltd now abandoned.
`
`FIELD OF THE INVENTION
`
`This invention relates generally to digital receivers
`for pseudorandom noise (PRN) encoded signals such as
`those used in passive ranging systems, and in particular
`to such a receiver that has been adapted for use in sig-
`naling environments susceptible to multipath fading.
`BACKGROUND OF THE INVENTION
`
`Passive ranging systems such as the United States’
`Global Positioning System (GPS) and the Russian
`Global Navigation System (GLONASS) allow a user to
`precisely determine latitude, longitude, elevation, and
`time of day. Ranging system receivers typically accom-
`plish this by decoding several precisely-timed signals
`transmitted by a group of special satellites.
`For example, within the GPS system, each signal
`transmitted by a satellite is modulated with low fre-
`quency (typically 50 Hz) digital data which indicates
`the satellite’s position and time of day, normalized to
`Greenwich Mean Time. Each satellite signal is further
`modulated with a unique, high frequency pseudoran-
`dom noise (PRN) code, which provide a mechanism to
`precisely determine the line of sight signal transmission
`time from each satellite.
`The GPS system satellite constellation has been
`placed in geostationary orbit such that at least four
`satellites are within a direct line of sight at any given
`position on the earth. A typical PRN receiver thus
`receives a composite signal consisting of several signals
`transmitted by the satellites, as well as any noise and
`interfering signals. A decoder or channel circuit may
`then recover one of the transmitted signals by correlat-
`ing (multiplying) the composite signal with a locally
`generated reference version of the PRN code signal
`assigned to the particular satellite of interest. If the
`locally generated PRN reference signal
`is properly
`timed, the digital data from that particular satellite may
`then be properly detected.
`The signals received from different satellites are also
`automatically separated by the multiplying process,
`because the signals transmitted by different satellites use
`unique PRN codes having low or zero cross-correlation
`power. The three dimensional position of the receiver
`and its velocity may then be resolved by using the PRN
`code phase information to precisely determine the
`transmission time from at least four satellites, and by
`detecting each satellite’s ephemeris and time of day
`data.
`
`In order to correctly determine the offset of the PRN
`reference signal, its relative time delay is typically var-
`ied relative to the incoming signal until a maximum
`power level in the resulting correlation signal is deter-
`mined. At the time offset corresponding to this point of
`
`5
`
`10
`
`15
`
`20
`
`25
`
`30
`
`35
`
`45
`
`50
`
`55
`
`60
`
`65
`
`2
`maximum received power, the local reference signal is
`considered to be in synchronism with the incoming
`signal, and the range measurement may then be made. A
`so-called delay lock loop (DLL) tracking system which
`correlates early, punctual, and late versions of the 10-
`cally generated PRN code signal against the received
`composite signal
`thus performs these operations to
`maintain PRN code lock in each channel.
`Because of this need to precisely determine the exact
`propagation time, a number of problems face the design-
`ers of PRN receivers. One problem concerns accurate
`phase and frequency tracking of the received signals;
`another problem concerns the correction of relative
`divergence between the received signals and the local
`PRN code signal generators in the presence of iono-
`spheric distortion.
`In addition, because GPS systems depend upon direct
`line of sight for communication propagation, any multi-
`path fading can further distort received signal timing
`estimates. In the ideal system, only one signal, the signal
`taking the direct or shortest path, is present. However,
`since the transmitter uses an omnidirectional (wide an-
`gle) antenna for maximum coverage, and since it is so
`far away from the receiver, the presence of surrounding
`reflecting objects such as buildings and natural surface
`formations means that there are typically multiple paths
`for the signal to take. Such a multipath signal takes a
`slightly different and longer router and thus arrives at
`the receiver at a different time.
`The exact number of multipath signals present at any
`given moment is a function of the satellite and antenna
`positions relative to any and all reflecting objects.
`Therefore, in the typical situation, there may be none,
`or there may be many multipath signals. Since the multi-
`path signals travel a longer distance they will always be
`received at some time after the direct path signal and
`will inevitably suffer a loss in power due to the reflec-
`tion(s). This time delay equals the difference in length
`between the direct path and the reflected path divided
`by the propagation velocity.
`The effect of the presence of multipath on the process
`of acquiring code lock is that there will always be some
`correlation with the multipath signals as well as with
`the desired, direct path signal.
`The typical way of dealing with this is to design the
`PRN autocorrelation functions such that even a small
`offset from zero in time will yield a near zero value in
`the estimate of the autocorrelation function. In reality,
`however, the autocorrelation power decreases linearly
`as the time offset increases,
`in either the position or
`negative directions The multipath correlation power
`only reaches zero when the PRN code offset is greater
`than plus or minus one chip. Since the carrier sampling
`rate is usually much higher than the PRN code chipping
`rate, partial correlations will occur at sub-chip offsets.
`Thus, in the presence of multipath distortion, most
`GPS receivers suffer a degradation in accuracy and an
`increase in processing time. This is especially true in
`high accuracy differential GPS applications, where
`pseudorange multipath will result in errors creeping
`into the differential corrections, causing large position
`biases.
`
`Unlike other error sources, multipath is typically
`uncorrelated between antenna locations. Thus, the base
`and remote receivers experience different multipath
`interference and as a result, simple differencing between
`them will not cancel the errors due to multipath distor-
`
`PETITIONERS 1004-0012
`
`

`
`5,414,729
`
`3
`tion. Also, modelling multipath for each antenna loca-
`tion is difficult and impractical.
`A common method of reducing multipath is to care-
`fully choose the design of the antenna and careful site
`selection. Unfortunately,
`it
`is often not possible to
`change either of these parameters. For example, if the
`antenna is to be mounted on an airplane fuselage, it will
`not be easily moved or replaced, and its shape is reces-
`sively restricted due to aerodynamic considerations.
`What is needed is a way to reduce the tracking errors
`present in PRN ranging receivers, especially those of
`the lower-frequency C/A code type, in the presence of
`multipath fading, without degrading the signal acquisi-
`tion capability of the receiver, or increasing errors due
`to Doppler shift, sudden receiver motion, or other noise
`sources. The desired method of reducing multipath
`distortion would be transparent to user, and operate
`within the GPS receiver itself, as opposed to requiring
`a special antenna or receiver siting.
`SUMMARY OF THE INVENTION
`
`4
`FIG. 1 is a high level block diagram of a PRN re-
`ceiver which incorporates the invention, including its
`downconverter, sampler, channel, and processor cir-
`cuits;
`FIG. 2 is a block diagram of one of the channel cir-
`cuits, showing multiple correlators being used in each
`channel;
`FIG. 3 is a block diagram of a carrier/code synchro-
`nizing circuit used in each channel circuit;
`FIG. 4 is a block diagram of a correlator circuit used
`in each channel circuit;
`FIG. 5 is a timing diagram showing the relative dura-
`tion of various portions of a received PRN signal;
`FIG. 6 is a plot of a direct path signal correlation, a
`multipath signal correlation, and the resulting direct
`with multipath correlation;
`FIG. 7 is a plot of another multipath correlation with
`different phase offset;
`FIG. 8 is a plot showing the nadlimiting effect of the
`channel 22;
`FIG. 9 shows the resulting tracking error; and
`FIG. 10 is another plot of the band limited correlation
`function showing the distribution of the multiple corre-
`lators.
`
`DETAILED DESCRIPTION OF A PREFERRED
`EMBODIMENT
`
`Now turning attention to the drawings, FIG. 1 is an
`overall block diagram of a pseudorandom noise (PRN)
`ranging receiver 10 constructed in accordance with the
`invention. It includes an antenna 11, a downconverter
`12, an in-phase (I) and quadrature (Q) sampler 14, a
`processor 16, a control bus 18, a channel bus 20 and
`multiple channels 22a. 22b, .
`.
`.
`, 22n, (collectively, the
`channels 22). The illustrated receiver 10 will be de-
`scribed herein as operating within the United States’
`Global Positioning System (GPS) using the commer-
`cial, coarse acquisition (C/A) pseudorandom codes,
`however, adaptations to other ranging systems are also
`possible.
`The antenna 11 receives a composite signal C3 consist-
`ing of the signals transmitted from all participating
`satellites within view, that is, within a direct line of sight
`of the antenna 11. When the GPS system is fully opera-
`tional, signals from at least four and as many as eleven
`satellites may be received simultaneously at each loca-
`tion on the earth.
`
`The composite signal C3 is forwarded to the down-
`converter 12 to provide an intermediate frequency sig-
`nal, IF. The IF signal is a downconverted and filtered
`version of the composite signal C5. The downconverter
`12 should have a bandpass filter which is sufficiently
`wide to permit several chips of the PRN coded signals
`to pass through. For the C/A code embodiment de-
`scribed here, this bandwidth is typically 8 MHz.
`The downconverter 12 also generates a sample clock
`signal, Fs, which is four times the frequency of the IF
`signal, which indicates the points in time at which sam-
`ples of the IF signal are to be taken by the sampler 14.
`The sampler 14 receives the IF and F5 signals and
`provides digital samples of the IF signal to the channels
`22 via the channel bus 20. The samples consist of in-
`phase (I) and quadrature (Q) samples of the IF signal
`taken at the times indicated by the Fs signal, typically
`by an analog-to-digital converter which samples at pre-
`cisely 90° phase rotations of the IF signal’s carrier fre-
`quency. With the digital sample clock signal, F5, chosen
`according to these guidelines, that is, with four samples
`
`PETITIONERS 1004-0013
`
`10
`
`15
`
`20
`
`25
`
`30
`
`Briefly, the invention is an improved receiver for
`pseudorandom noise (PRN) encoded signals consisting
`of a sampling circuit, multiple carrier and code synchro-
`nizing circuits, and multiple correlators, with each cor-
`relator having a selectable code delay spacing. The time
`delay spacing of the multiple correlators is distributed
`around an expected correlation peak to produce an
`estimate of the correlation function parameters which
`vary with respect to multipath distortion. The parame-
`ters of interest discemable from an estimate of the shape
`of the autocorrelation peak include the direct signal
`path time and phase offsets.
`This information may be used in turn used to deter-
`mine the offset estimates for locally generated PRN
`reference code and carrier phase tracking signals, or
`may be used to adjust a range measurement.
`If multipath correlators are not available full-time for
`each channel, then the code delay for a fewer number of
`correlators, such as two for each channel, can be se- 40
`quenced from epoch to epoch, so that over time, mea-
`surements from several points on the correlation func-
`tion can be taken.
`In another embodiment, the majority of the channels
`in a receiver can be left to operate normally, with one or
`more of the channels being dedicated to continuously
`sequencing from channel to channel to determine the
`multipath parameters for a partial PRN code being
`tracked.
`
`35
`
`45
`
`There are several advantages to this arrangement. In
`environments such as commercial GPS coarse/acquisi-
`tion (C/A) code applications, where the multipath dis-
`tortion in the received composite signal is of the same
`order of magnitude as a PRN code chip time, the PRN
`receiver is capable of acquiring carrier and code lock
`over a wide range of operating conditions. Once the
`receiver is locked, it will automatically remain locked,
`even in the presence of multipath distortion.
`An improvement in range measurement accuracy is
`actively accomplished without special antenna design,
`and without specifying the particular location of the
`antenna, even in differential ranging applications.
`BRIEF DESCRIPTION OF THE DRAWINGS
`
`The above and further advantages of the invention
`may be better understood by referring to the following
`description in conjunction with the accompanying
`drawings, in which:
`
`50
`
`55
`
`60
`
`65
`
`

`
`5,414,729
`
`5
`taken in every IF carrier cycle, the output samples from
`the sampler 14 are thus in in-phase and quadrature order
`as I,Q, -I,-Q,I,Q .
`.
`. and so on. The I and Q samples are
`then separated and routed to the channels 22 on sepa-
`rate 15 and Q5 conductors of the channel bus 20, along
`with the F5 signal. For more details of one embodiment
`of the downconverter 12 and sample 14, please refer to
`U.S. Pat. No. 5, 101,416 entitled “Multichannel Digital
`Receiver for Global Positioning System” issued Mar.
`31, 1992, and assigned to NovAtel Communications
`Ltd.
`Each channel 22 is normally assigned to process the
`signal transmitted by one of the satellites which is pres-
`ently within view of the antenna 11. A given channel 22
`thus processes the Is and Q5 signals and tracks the carrier
`and code of the signal transmitted by its assigned satel-
`lite. As explained below, each charmel 22 uses a carrier
`and code synchronizing circuit to frequency and phase-
`track the PRN encoded carrier signal by maintaining an
`expected Doppler offset unique to the desired satellite.
`Furthermore, each charmel 22 contains multiple corre-
`lators to maintain phase lock with a locally generated
`PRN code reference signal as well, to remove the ef-
`fects of any multipath distortion on the position mea-
`surement.
`
`The locally generated PRN code reference signal is
`then used to decode the data from the assigned satellite.
`The resulting decoded data,
`including the satellite’s
`ephemeris, time of day, and status information, as well
`as the locally generated PRN code phase and carrier
`phase measurements, are provided to the processor 16
`via the control bus 18. The channels 22 are described in
`detail in connection with FIG. 2.
`
`The sampler 14 and channels 22 are controlled by the
`processor 16 via the control bus 18. The processor 16
`includes a central processing unit (CPU) 162 which
`typically supports both synchronous-type input/output
`(I/O) via a multiple-bit data bus DATA, address bus
`ADDR, and control signals CTRL and synchronous
`controller circuit 164, and an interrupt-type 1/0 via the
`interrupt signals, INT and an interrupt controller circuit
`166. A timer 168 provides certain tinting signals such as
`a measurement trigger MEAS indicating a request for a
`range measurement to be taken. The operation of the
`processor 16 and its various functions implemented in
`software will be understood from the following discus-
`sron.
`
`The composite signal C; received from the antenna 11
`typically consists of signals transmitted by all satellites
`within view (that is, within a direct line-of-sight of the
`receiver 10), any interfering signals, such as multipath
`signals and noise. The carrier signal used by the GPS
`C/A ranging system is an L-band carrier at 1.57542
`GigaHertz (GHz) with a PRN code rate of 1.023 MHz
`and a nominal transmitted power of -160 dBW. Natu-
`ral background noise at about -204 dBW/Hz is typi-
`cally mixed in with the L-band signals. In addition, one
`or more multipath signals are present in the composite
`signal C3, as will be described below. For more detailed
`information on the format of the GPS system signals,
`see “Interface Control Document ICD-GPS-200, Sept.
`26, l984”, published by Rockwell International Corpo-
`ration, Satellite Systems Division, Downey, Calif.
`90241.
`FIG. 5 shows, on a distorted time scale, the relative
`durations of various components of a typical PRN rang-
`ing signal transmitted by a GPS satellite and certain of
`the signals in a preferred embodiment of a channel 2272.
`
`6
`A single carrier cycle has a particular duration, C. A
`single cycle of the digital sample signal clock F5, con-
`sists of K carrier cycles.
`A PRN code chip includes N cycles of the F3 signal,
`and a PRN code epoch consists of Z PRN code chips,
`where Z is also known as the sequence length of the
`PRN code. One data bit typically consists of T PRN
`code epochs. For the preferred embodiment of the in-
`vention adapted to receive the GPS L 1 ranging signal,
`the carrier frequency is 1575.42 MHZ, and K is 77, so
`that F, equals 20.46 MHz. In addition, a constant, N is
`20, so that the PRN code chip rate is 1.023 MHz, and Z
`is 1023, so that the PRN code epoch rate is 1 kHz. T,
`another constant, is also 20, so that the data bit rate is 50
`Hz.
`A channel circuit 2222 is shown in detail in FIG. 2. It
`includes a carrier and code synchronizer circuit 220, a
`PRN code generator 230, a carrier phase shifter 235,
`multiple correlators 240-1,240-2, 240-3 .
`.
`. 240—m (col-
`lectively, correlators 240), a code shift register 250, an
`early minus late (or early, late) discriminator 260, a dot
`product (or punctual, early minus late) discriminator
`265, and a multipath parameter estimator 270.
`The PRN code generator 230 uses signals output by
`the synchronizer 220 to generate a local PRN reference
`signal, PRN CODE. The particular PRN CODE signal
`generated at any given time depends upon which satel-
`lite it is desired for the channel to be tuned to, as se-
`lected by the SAT ID input. PRN code generators such
`as code generator 230 are well known in the art.
`The synchronizer 220 is a single numerically con-
`trolled oscillator (NCO) which uses the sample clock
`F5 and appropriate instructions from the processor 16 to
`provide the control signals required by PRN code gen-
`erator 230 and correlators 240 in the carrier or code
`carrier phase to non-coherently track the errors caused
`by residual Doppler and multipath distortion.
`Before continuing with a detailed discussion of the
`charmel 2222, refer to FIG. 3, which is a detailed block
`diagram of the carrier and code synchronizer 220. This
`element includes an expected Doppler rate register 221,
`an accumulated delta range (ADR) register 222, and a
`fine chip counter 224. A code phase generator circuit
`226 also includes a subchip counter 226a, chip counter
`226b, epoch counter 226d. Buffer circuits 227, 228, and
`229 allow the processor 16 to load, read, and add to or
`subtract from the contents of the various counters and
`registers in the synchronizer 220.
`The synchronizer 220 accepts the sample clock signal
`F; from the charmel bus 20, an expected Doppler value
`EDOPP and corrected values for the registers and
`counters 222, 224 and 226 from the control bus 18. In
`response to these inputs, it provides a clock signal CLK
`and reset signal RST to the PRN code generator 230, as
`well as interrupt signals INTI, INT4, and INT20 to the
`control bus 18. An instantaneous carrier phase angle
`estimate is also provided via bits wro, m, .
`.
`. 7rp to the
`correlators 240.
`The contents of the ADR register 222 and code phase
`generator 226 provide an instantaneous estimate of the
`transmit time of the particular satellite signal assigned to
`charmel 22n. The difference between this estimate of the
`transmit time and the receiver time of day (as estimated
`by the timer 168 in FIG. 1) is then taken as the propaga-
`tion time of that signal plus any receiver clock offset. By
`multiplying the propagation time by the speed of light,
`a precise measurement of the range from the receiver 10
`to the assigned satellite may then be made by the pro-
`
`10
`
`15
`
`20
`
`25
`
`30
`
`35
`
`45
`
`50
`
`55
`
`60
`
`65
`
`PETITIONERS 1004-0014
`
`

`
`5,414,729
`
`8
`accomplished non-coherently, in the sense that the local
`reference signal, Fs, need not remain phase locked to
`the carrier of the intermediate frequency signal, IF, in
`order for the PRN code generator 230 to remain phase-
`locked.
`
`7
`cessor 16. These measurements occur at selected time
`indicated by the measurement strobe MEAS from the
`timer 168, and are typically taken simultaneously across
`all the channels 22. The resulting range to each satellite
`is then used by the processor 16 to compute the position
`and or velocity of the receiver 10.
`In operation, the expected Doppler rate register 221
`is first loaded via the processor bus 18 with an estimated
`Doppler offset EDOPP for the particular satellite
`tracked by channel 22n. In most instances, such as when
`the receiver 10 has been operating for some time, the
`EDOPP estimate may be taken from almanac data al-
`ready received from satellites to which the receiver 10
`has been synchronized, since the almanac data from
`each satellite includes an estimated position of all other
`operating satellites. However, if this almanac data is not
`available, such as when the receiver 10 is first turned on,
`this estimate may be determined by successive approxi-
`mation techniques, as described in the aforementioned
`US. Pat. No. 5,101,416.
`The Doppler value is specified in carrier Doppler
`cycles per F5 pulse. For example, if the expected Dop-
`pler frequency is +4.45 kilohertz (kHz), which is a
`possible Doppler frequency for a stationary receiver
`and an approaching satellite, dividing by a typical F5
`frequency of 20.46 MHz for the GPS L1 embodiment
`results in an expected Doppler shift of approximately
`0.00044 carrier cycles per F5 pulse. Specified in this
`way, the Doppler value will always be less than one.
`The ADR register 222, which provides an estimate of 30
`the range to the satellite being tracked, is divided into a
`whole cycle portion 222w and a partial cycle portion
`222p. As shown, an adder 223 is arranged to add the
`contents of the Doppler register 221 to the partial cycle
`portion 222p of the ADR 222 upon the occurrence of 35
`every Fs pulse. The most significant bits 0, 1, .
`.
`.,pof
`the partial cycle portion 222p thus provides an instanta-
`neous expected carrier phase angle in cycles.
`When the partial cycle register 222p has a carry out,
`the whole number portion 222w is incremented and the
`fine chip counter 224 is also incremented. If the partial
`cycle register 222p requires a borrow, then the whole
`number portion 222w and fine chip counter 224 are
`decremented.
`
`The most significant bit of the subchip counter 226:: is
`used as a clock signal, CLK, to indicate a PRN code
`chip edge. In the preferred embodiment for the GPS L
`1 carrier, the subchip counter 226a counts from zero to
`nineteen since N equals twenty, i.e., there are twenty F5
`cycles per PRN code chip (FIG. 5).
`The chip counter 22617 is used to determine the dura-
`tion of a complete PRN code sequence. For the GPS
`embodiment, there are 1,023 C/A code chips in a PRN
`code epoch, and thus the chip counter 226b counts from
`zero to 1022. The most significant bit, INTI indicates
`the end of a complete PRN code epoch to the processor
`16; it is also used to reset the local PRN code generator
`230. Another clock signal, INT4, which is four times
`the rate of INT1 (i.e., the third most significant bit of
`the chip counter 226b) is also generated. Both INT1 and
`INT4 may be used to interrupt the processor 16 to
`service the correlators 240 during an initial
`locking
`sequence, as will be described shortly.
`Finally, the epoch counter 226d is used to indicate the
`end of a data bit, after T PRN code epochs. This indica-
`tion is given by the most significant bit of the epoch
`counter 226d, which is output as the INT20 signal.
`The carrier tracking loop is inherently much more
`sensitive than the code DLL and is able to measure
`' small changes extremely accurately. Assuming the car-
`rier loop is tracking properly, the fine chip counter 224
`in conjunction with the subchip counter 226a, enables
`the channel 22n to accurately track any relative motion
`of the receiver 10 with respect to the satellite.
`By returning now to FIG. 2, the operation of a typi-
`cal receiver channel 22n will be understood in greater
`detail. A carrier phase shifter 235 accepts the I; and Q5
`samples, along with the instantaneous carrier phase bits
`771, 7r2, .
`.
`. , 771,, from the synchronizer 220. The carrier
`phase shifter 235 then phase rotates the Is and Q5 sam-
`ples by an amount indicated by the instantaneous carrier
`phase angle estimate generated by the synchronizer 220,
`and provides outputs ID and Qp according to the fol-
`lowing expressions:
`
`ID=I; cos(II)+ Q, sin(II)
`
`Q1): Q5 cos(II)—I5 sin(II)
`
`where PRN is the current value of the PRN CODE
`input and II is the instantaneous carrier phase estimate
`represented by the bits 771, 712,
`.
`.
`.
`, 771,. Since the Is and
`Q5 samples are in digital form, this phase shift operation
`is performed by appropriate digital circuits. By remov-
`ing the instantaneous carrier shift in the same operation
`at every F5 clock pulse, signals with very high Doppler
`offsets may be processed before any significant power
`loss is encountered.
`
`The shift register 250 receives the PRN code signal
`from the PRN code generator 230 and generates a plu-
`rality, m, of time-shifted PRN code replica signals
`PRNt1, PRNt2, . . .
`, PRNtm. The time shifts imparted to
`the PRN code replica signals are typically distributed
`about an expected code time delay, in a manner which
`will be described below in connection with FIGS. 6
`through 10.
`
`PETITIONERS 1004-0015
`
`10
`
`15
`
`20
`
`25
`
`The subchip counter 226a is clocked by the F5 signal
`and controlled by the fine chip counter 224. Subchip
`counter 226a is nominally a O to N-l counter controlled
`directly by the Fs signal, but may be adjusted to count
`one extra cycle or one fewer cycle depending upon the
`state of the fine chip counter 224. In particular, when
`the fine chip counter carries out, i.e., increments from
`K-1 to O, a cycle is stolen from the sub chip counter
`226a to keep it synchronized with the ADR 222. In
`other words, this event causes the subchip counter 226a
`to count only to N——2 for one iteration.
`When th