US005929806A
`5,929,806
`(114) Patent Number:
`United States Patent 55
`Birchler et al.
`[45] Date of Patent:
`Jul. 27, 1999
`
`
`[54] METHOD FOR ESTIMATING A LOCATION
`OF A MOBILE UNIT BASED ON AT LEAST
`TWO FIXED TRANSCEIVERS
`
`[75]
`
`Inventors: MarkA.Birchler, Roselle; Debra A.
`Jones; Nicholas C. Oros, both of
`Schaumburg,all of IIL.
`
`[73] Assignee: Motorola, Inc., Schaumburg,Ill.
`
`[21] Appl. No.: 08/846,751
`
`[22]
`
`Filed:
`
`Apr. 30, 1997
`
`Int. Che ees GO1S 5/02; HO4B 7/185
`[51]
`[52] U.S. Che occ
`eeceeeeseeseenteees 342/357; 701/213
`[58] Field of Search oe 342/357, 457;
`701/213
`
`[56]
`
`References Cited
`U.S. PATENT DOCUMENTS
`
`5,202,829
`5,416,712
`5,436,632
`
`4/1993 Geier.
`5/1995 Geieret al.
`7/1995 Sheynblat .
`
`.
`
`5,525,998
`5,736,960
`
`6/1996 Geier .
`4/1998 Murphy et al. oo. 342/357
`
`Primary Examiner—Mark Hellner
`Attorney, Agent, or Firm—Kenneth A. Haas
`
`[57]
`
`ABSTRACT
`
`Signals comprising time of transmission indicationsrelative
`to a commontime base (302) are transmitted by at least two
`fixed transceivers (110-116). When the signals are received
`by a mobile unit (160), a time of reception indicationrelative
`to the common time base is calculated for each signal.
`Pseudo-ranges are calculated based on the time of transmis-
`sion and time of reception indications, and a WLS solution
`location estimate is calculated based on the pseudo-ranges.
`Essentially the same procedure may be performed on signals
`transmitted by the mobile unit. Given the commontime base
`between the fixed transceivers and the mobile unit, only two
`or more signals are required to determine the location
`estimate. Where GPS receivers are used to supply the
`common time base,
`location determinations can still be
`made when a GPSsolution is unavailable.
`
`15 Claims, 3 Drawing Sheets
`
`140
`
`
`
`INFRASTRUCTURE
`ENTITY
`
` LOCATION
`PROCESSOR
`130
`
`APPLE 1021
`
`APPLE 1021
`
`1
`
`

`

`U.S. Patent
`
`Jul. 27, 1999
`
`Sheet 1 of 3
`
`5,929,806
`
`
`
`INFRASTRUCTURE
`ENTITY
`
`
`LOCATION
`
`PROCESSOR
`
`130
`
`100
`
`
`
`
`302 utanee UNIT/FIXED[-30!
`TRANSCEIVER
`
`
`|
`
`
`
`
`
`MOBILE
`
`2
`
`

`

`U.S. Patent
`
`Jul. 27, 1999
`
`Sheet 2 of 3
`
`5,929,806
`
`
`
`
`
`RECEIVE OUTBOUND SIGNALS
`TRANSMITTED RELATIVE 10
`COMMON TIME BASE
`
`
`
`
`
`DETERMINE TIME OF RECEPTION
`
`FOR EACH SIGNAL RELATIVE 10
`COMMON TIME BASE
`
`DETERMINE PSEUDO-RANGE FOR
`
`| TRANSMIT LOCATION ESTIMATE |
`|
`TO MOBILE UNIT/
`|
`INFRASTRUCTURE ENTITY
`Ll. 2 eS ee ae ae J
`
`
`
`
`
`;SENDTINEOF TRANSMISSION |
`
`;
`AND RECEPTION INDICATIONS
` DETERMINE TIME OF RECEPTION
`
`__10 LOCATION PROCESSOR
`
`FOR EACH SIGNAL RELATIVE TO
`
`
`COMMON TIME BASE
`EACH SIGNAL
`
`
`
`SEND TIME OF TRANSMISSION
`AND RECEPTION INDICATIONS
`TO LOCATION PROCESSOR
`
`
`
`
`DETERMINE PSEUDO-RANGE FOR
`
`
`
`RECEIVE INBOUND SIGNALS
`TRANSMITTED RELATIVE TO
`COMMON TIME BASE
`
`EACH SIGNAL
`
`T FAdueur?T TAALTTAL Py S— -- ol
`TRANSIT ate ESTIMATE
`| INFRASTRUCTURE ENTITY __ |
`
`3
`
`

`

`U.S. Patent
`
`Jul. 27, 1999
`
`Sheet 3 of 3
`
`5,929,806
`
`WEIGHTED
`LEAST-SQUARES SOLUTION
`
`DETERWINE Aj AND C BASED ON
`NOMINAL LOCATION ESTIMATE
`
`CALCULATE Ayy
`
`504
`
`506
`
`UPDATE NOMINAL LOCATION
`ESTIMATE, Ap,
`AND C BASED ON Ayy
`
`
`508
`
`
`Ayy YES
`
`
`DIVERGING
`
`?
`
` END
` Ayy
`9
`
`CONVERGED BEYOND
`LOWER THRESHOLD(S)
`
`YES
`
`SET LOCATION EQUAL 10
`NOMINAL LOCATION ESTIMATE
`
`514
`
`FIG. 5
`
`4
`
`

`

`5,929,806
`
`1
`METHOD FOR ESTIMATING A LOCATION
`OF A MOBILE UNIT BASED ON AT LEAST
`TWO FIXED TRANSCEIVERS
`
`FIELD OF THE INVENTION
`
`The present invention relates generally to location deter-
`mination in wireless communication systems and,
`in
`particular, to a method for estimating a location based on at
`least two fixed transceivers.
`
`BACKGROUND OF THE INVENTION
`
`Wireless communication systems are known in which
`mobile units (e.g., in-car mobile or in-hand portable radios)
`wirelessly communicate with a fixed communication infra-
`structure comprising a plurality of geographically-diverse
`transceivers.
`In such systems, methods for determining
`location information for a given mobile unit are known. In
`particular, the well-known weighted least squares (WLS)
`solution can be used to determine location information as
`
`shown, for example, in U.S. Pat. No. 5,416,712 issued to
`Geieret al.
`
`Succinctly stated, the WLS approachto location determi-
`nation attempts to iteratively derive a location estimate for
`a mobile unit based, in part, on distance estimates between
`the mobile unit and fixed transmitters having known loca-
`tions. Given that distance can be calculated as the product of
`velocity and time,
`the distance estimates (referred to as
`pseudo-ranges or PRs) are calculated in practice by multi-
`plying the propagation delays between the mobile unit and
`fixed transmitters with the speed of light. Assuming ideally
`measured propagation delays, the location of the mobile unit
`can be calculated using the pseudo-ranges with little or no
`error. However, propagation delays are measured in practice
`using transmitted signals, which signals are subject to the
`effects of various error sources, e.g., noise, multipath
`interference, distortion, etc. The resultant errors in the delay
`measurementsare translated into errors in the pseudo-ranges
`and, consequently,
`into error in the location estimate. In
`order to combat the presence of measurementerrors, the
`WLSsolution factors the reliability of the various measure-
`ments into the location estimation.
`
`Generally, location determinations in three-dimensional
`space require the reception of signals transmitted by at least
`four fixed transmitters. For example,
`in the well-known
`Global Positioning System (GPS), the receiving unit (i.e.,
`the unit for which a location is to be determined) does not
`maintain a timing reference exactly synchronized to the
`highly stable and synchronized timing references main-
`tained by the transmitting satellites. As a result, the pseudo-
`ranges determined by the receiving unit will not result in a
`precise location solution. In order to combat this, a fourth
`measurementis used to indicate exactly how far out of synch
`the receiving unit’s local time base is with respect to the
`universal time base maintained by the satellites. In this
`manner, the receiving unit can determine a correction factor
`that, when applied to the pseudo-ranges, results in a more
`precise location solution. Additionally, by applying the
`correction factor to the local time base, the receiving unit
`can, at least for a period of time, serve as an accurate time
`reference relative to the universal time base.
`
`This same concept may be readily applied to two-
`dimensional problemsas well, e.g., land mobile radio com-
`munication systems where altitude measurements are of
`lesser importance. However, the same difficulty regarding
`synchronized time basesis also present in two-dimensional
`applications. It is knownin theart to provide a commontime
`
`10
`
`15
`
`25
`
`30
`
`35
`
`40
`
`45
`
`50
`
`55
`
`60
`
`65
`
`2
`base amongfixed transceivers in a land mobile radio system.
`Indeed, such a common timebaseis often provided through
`the use of GPSreceivers, described above. Additionally, as
`GPS receivers become increasingly less expensive,
`they
`may be more readily incorporated into mobile and portable
`units. Future developments may also provide for highly
`accurate time synchronization between fixed transmitters
`and mobile units. As such, a need exists for a method which
`incorporates the advantages of a common time base into a
`WLSlocation solution.
`
`BRIEF DESCRIPTION OF THE DRAWINGS
`
`FIG. 1 is a block diagram of a wireless communication
`system in accordance with the present invention.
`FIG. 2 is a flowchartillustrating a method for determining
`a location estimate based on outboundsignals in accordance
`with the present invention.
`FIG. 3 is a block diagram of a mobile unit or a fixed
`transceiver in accordance with the present invention.
`FIG. 4 is a flowchart illustrating a method for determining
`a location estimate based on inbound signals in accordance
`with the present invention.
`FIG. 5 is a flowchart illustrating the determination of a
`location estimate using a weighted least squares solution.
`
`DESCRIPTION OF A PREFERRED
`EMBODIMENT
`
`The present invention provides a method for estimating a
`location of a mobile unit
`in a wireless communication
`system using a common time base between the mobile unit
`and fixed transceivers.
`In one embodiment, signals are
`transmitted by at least two fixed transceivers, each of the
`signals including a time of transmission indication generated
`by a corresponding fixed transceiver. When the signals are
`received by a mobile unit, a time of reception indication
`relative to the common time base is calculated for each
`signal. Pseudo-ranges corresponding to each signal are cal-
`culated based on their respective time of transmission and
`time of reception indications. Subsequently, a WLSsolution
`location estimate is calculated based on the pseudo-ranges
`thus determined. In a second embodiment, essentially the
`same steps are performed relative to inbound signals (i.e.,
`signals transmitted by the mobile unit). Because there is a
`common time base between the fixed transceivers and the
`
`mobile unit, only two or more signals are required to
`determine the location estimate. In GPS equipped systems,
`such a method can be advantageously employed to provide
`location determination when insufficient data is available to
`provide a GPS-based location.
`The present invention can be morereadily described with
`reference to FIGS. 1-5. FIG. 1 is a block diagram of a
`wireless communication system 100, such as an IDEN™
`system manufactured by Motorola, Inc. The wireless com-
`munication system 100 comprises geographically-diverse
`fixed transceivers 110-116 coupled to a location processor
`130 via a network 140, and a plurality of mobile units 160
`(only one shown). Although not shownin FIG. 1,each fixed
`transceiver 110-116 and mobile unit 160 includes memory
`and one or more digital computing devices used to store and
`execute software programs, as knownin the art.
`Additionally,
`the fixed transceivers 110-116 and the
`mobile units 160 each have access to a commontimebase,
`as illustrated in FIG. 3. As shown in FIG. 3, a given fixed
`transceiver/mobile unit 301 is coupled directly to a common
`time base 302, although in practice the common time base
`
`5
`
`5
`
`

`

`5,929,806
`
`3
`may form an integral part of the fixed transceiver/mobile
`unit 301. In the preferred embodiment, the timing reference
`provided by each commontimebase is stable on the order
`of 100 nanoseconds or less, thereby maintaining accurate
`time base synchronization throughout the system. A suitable
`commontime base 302is available using the corrected clock
`provided by GPSreceivers. Of course, if a GPS receiver is
`available at a mobile unit, location determinations can be
`readily made without
`the use of the present
`invention.
`However, the present invention may be usefully employed in
`those situations in which the GPS receiver is unable to
`provide a location (i.e.,
`the GPS receiver is not able to
`acquire enough satellite signals for an accurate location
`determination), but where the timing reference provided by
`the GPSreceiverisstill available for use. The reliability of
`a timing reference provided by a GPSreceiver is dependent
`upon the quality of the local oscillator(s) used to maintain
`the timing reference. Thus, during periods of GPS
`unavailability, the length of time during which the timing
`reference could be reliably used is dependent upon the
`stability of the local oscillator. However, knowing the sta-
`bility of the local oscillator, it is possible to determine the
`maximum period during which the local oscillator may be
`relied upon to provide an accurate timing reference before
`additional corrections will be required. Additionally, it may
`be possible to model the known performance of the local
`oscillator in order to perform periodic corrections.It is also
`anticipated that technology could advanceto the point where
`highly stable timing mechanisms requiring only occasional
`synchronization may become available. Regardless,
`the
`availability of the common time base 302 allows both the
`mobile units and fixed transceivers to reliably compare
`transmission and reception times for signals transmitted
`within the system 100.
`The coverage area provided by each transceiver 110-116
`gives rise to a corresponding cell 120-126, represented by
`hexagons as shown. In practice, each transceiver 110-116
`may actually comprise a bank of transceivers, although it is
`assumed hereinafter that each cell includes only a single
`transceiver for simplicity. While a mobile unit
`roams
`throughout the system 100, wireless communication ser-
`vices for the mobile unit are managed and provided by a
`serving site. Neighbor sites are those sites surrounding the
`current serving site. In the example shownin FIG. 1, the site
`identified labeled X, functions as the serving site for the
`mobile unit 160, and the sites labeled X,—X, correspond to
`neighborsites.
`The transceivers 110-116—which in the preferred
`embodiment provide wireless channels according to a time-
`division multiplexed (TDM) format—andthe location pro-
`cessor 130 collectively are referred to as a fixed infrastruc-
`ture. Being fixed, the precise location of each transceiver
`110-116 is known with high accuracy as determined by
`surveying or other suitable measurement techniques.
`As described below,the present inventionrelies on signals
`received or transmitted by a mobile unit. In the preferred
`embodiment, each of these signals comprise known data
`symbols unique to each transceiver. For example,
`in an
`IDEN™ system, each transceiver will transmit its known
`data symbols whenever an unused time slot becomesavail-
`able.
`The network 140 allows data and control information to
`be conveyed between the various infrastructure elements
`and operates in accordance with well-known network pro-
`tocols. The fixed infrastructure may also include various
`other network entities 170 that may serve as the requesters
`of, or the destination for, location information for a given
`
`10
`
`15
`
`20
`
`30
`
`35
`
`40
`
`45
`
`50
`
`55
`
`60
`
`65
`
`4
`mobile unit. Examples of such infrastructure entities
`include, but are not limited to, consoles, telephone intercon-
`nect devices, and management terminals.
`The location processor 130 includes a digital computing
`device 131 and memory 132 used for the storage and
`execution of software programs. In practice, a commercially
`available computer workstation or a customized computer
`platform may be used to implement the location processor
`130. In the preferred embodiment, the methods described
`hereinafter are implemented as software programs resident
`in the location processor 130 and/or mobile units 160.
`In the preferred embodiment, the location processor of
`FIG. 1 cooperates with a mobile unit when determining a
`location for that mobile unit, although it is also possible for
`the mobile unit
`to determine its own location without
`assistance from the location processor. FIG. 2 illustrates a
`flowchart for determining a location estimate of a mobile
`unit based on outboundsignals (i.e., signals transmitted by
`fixed transceivers). At step 202, the mobile unit receives
`signals transmitted by at least two geographically-diverse
`transceivers. Each of the signals uniquely corresponds to
`one of the transceivers, and each signal comprises known
`data symbols transmitted by its corresponding transceiver.
`Additionally, each of signals received at step 202 includes a
`time of transmission indication relative to the common time
`base.
`
`In the preferred embodiment, the time of transmission
`indication is inherently provided by virtue of the time slot
`structure in the system. For example, in an IDEN™system,
`a structure of sequentially numbered time slots is provided
`throughout
`the system. Because all
`transmitting entities
`throughout the system (i.e., mobiles and fixed transceivers)
`have access to the common time base, the integrity of the
`time slot structure is maintained with high accuracy. Any
`device wishing to transmit within the system can do so by
`transmitting during a time slot knownto be available for use.
`Additionally, repetition of the numbering sequence used to
`identify each of the time slots is on the order of many
`seconds, effectively eliminating any ambiguity as to when a
`time slot actually starts. By ascertaining the number of a
`given received timeslot, and given that all timeslots are of
`equal duration, the receiving entity (in this case, a mobile
`unit) is provided with information regarding when the time
`slot was transmitted. In effect, the time slot number func-
`tions as a time of transmission indication relative to the
`common time base.
`
`At step 204, the mobile unit determines time of reception
`indications for each of the signals received at step 202.
`Because each of the signals comprises a known sequenceof
`symbols unique to each transceiver, a receiving entity (in
`this case, a mobile unit) performs a running correlation
`between received symbols (i.e.,
`those symbols that have
`been demodulated) and the known sequence(s) of symbols.
`Whena high correlation is found, indicating that a known
`sequence of symbols has been received, the time is noted
`relative to which slot number(s) provided the relevant sym-
`bols. Because the time slot structure is maintained with high
`accuracy relative to the common time base,
`the time of
`reception can be determined. If the time of transmission
`indication is determined relative to the beginning of the
`known sequence of symbols,the time of reception indication
`is also based on when the first symbol of the known
`sequence of symbols wasreceived. Of course, other symbols
`within the known sequence of symbols could be used. Also,
`in order to determine exactly when the relevant symbols
`were received, compensations can be made to account for
`the amount of known processing delay in the receiver,e.g.,
`
`6
`
`6
`
`

`

`5,929,806
`
`5
`the amountof time necessary to demodulate the symbols and
`compute the correlations.
`Atstep 206, the mobile unit can optionally send the time
`of transmission and reception indications to the location
`processor. In practice, the information sent to the location
`processor would include identifications of the fixed trans-
`ceivers that transmitted the signals corresponding to the time
`of transmission and reception indications.
`The delay between when the known data symbols were
`transmitted and when they were received serve as the basis
`for determining pseudo-ranges between the mobile unit and
`each of the transceivers. At step 208,
`the pseudo-ranges
`corresponding to each signal received at step 202 is deter-
`mined. If step 206 was executed,
`the determination of
`pseudo-ranges at step 208 is performed by the location
`processor; if step 206 was not executed, step 208 is per-
`formed by the mobile unit. For each signal, the delay time
`delimited by the time of transmission indication and the time
`of reception indication (or an average delay time between
`multiple sets of time of transmission and reception indica-
`tions for known symbols within a given signal) is deter-
`mined. Assuming speed of light propagation (and possibly
`accounting for atmospheric delays or any other known
`delays), the pseudo-ranges are readily calculated from the
`delay times thus determined. Referring to the example
`shown in FIG. 1, the mobile unit 160 is assumed to receive
`signals from two transceivers,
`labeled X, and X,, each
`transceiver also having associated with it a pseudo-range Dy
`and D,, respectively. Although the serving site X, is used in
`the above example, the neighboring sites alone could just as
`easily be used.
`At step 210, having determined pseudo-ranges for each
`signal, either the location processor (if step 206 was
`executed) or the mobile unit Gif step 206 was not executed)
`determines a WLSsolution to the equation:
`
`VAp = VCR,
`
`The components of the above equation and the steps for
`determining the WLSlocation solution are further described
`with reference to FIG. 5 below. The result of determining the
`WLSlocation solution is a location estimate for the given
`mobile unit. At step 212, the location estimate is optionally
`sent to a mobile unit or infrastructure entity. That is, if the
`mobile unit performs step 210, the resulting location esti-
`mate can be sent to an infrastructure entity (or even the
`location processor) that may have requested the location
`information. Alternatively,
`if the location processor per-
`forms step 210,
`the location estimate can be sent to the
`mobile unit that performed steps 202-206, or to a different
`mobile unit, or to an infrastructure entity. Regardless, the
`method shown in FIG. 2 allowsfor reliable location deter-
`mination based on signals transmitted by two or more fixed
`transceivers, rather than a minimum ofthree as required in
`prior art solutions.
`Turning now to FIG. 4, a flowchart for determining a
`location estimate of a mobile unit based on inbound signals
`(i.c., signals transmitted by the mobile unit) is shown. At
`step 402, at least two geographically-diverse transceivers
`each receive a signal transmitted by a mobile unit. That is,
`each transceiver receives the same signal, but subject to
`different error conditions(i.e., fading, interference, etc.) than
`the other transceivers. For example, referring to FIG. 1,
`transceivers X, and X, receive a signal transmitted by the
`mobile unit 160. In the preferred embodiment, the signal
`transmitted by the mobile unit is a sequence of known data
`
`10
`
`15
`
`20
`
`25
`
`30
`
`35
`
`40
`
`45
`
`50
`
`55
`
`65
`
`7
`
`6
`symbols equivalent in operation to those transmitted by the
`transceivers, as described above. Additionally, the received
`signals at each transceiver include time of transmission
`indications relative to the common time base in the same
`
`manner as described above with respect to step 202.
`At step 404, each transceiver from step 402 determines
`time of reception indications for its corresponding received
`signals in a similar manner as described above with respect
`to step 204. At step 406, each transceiver sendsits respective
`time of transmission and reception indications to the loca-
`tion processor.
`In practice,
`the information sent
`to the
`location processor would includeidentifications of the fixed
`transceivers that received the signals corresponding to the
`time of transmission and reception indications.
`Atsteps 408 and 410, respectively, the location processor
`calculates the pseudo-ranges corresponding to each of the
`received signals and determines a WLSsolution (described
`further in FIG. 5 below) to provide the location estimate for
`the mobile unit. At step 412, the location estimate is option-
`ally sent
`to a mobile unit (either the mobile unit
`that
`transmitted the signals received at step 402 or a different
`mobile unit) or infrastructure entity.
`FIG. 5 illustrates a method for determining a location
`estimate using a WLSsolution. At step 502, a pseudo-range
`correction vector ne and a direction cosine matrix c¢ are
`determined based on a nominal location estimate, denoted
`by the coordinates (x,,y,,). The initial nominal
`location
`estimate is selected to be within the boundaries of a cell
`
`corresponding to one of the fixed transceivers transmitting
`or receiving the signals used in the location determination
`(preferably the cell corresponding to the servingsite), and is
`also selected based upon the type of cells being used. That
`is, if the cell is “sectorized”, as known in the art, the center
`of massof the serving sector is selected as the initial nominal
`location estimate. However, if the first cell is not sectorized,
`a good choice for an initial nominal location estimate is a
`point on a circle centered on, and equal to approximately
`half the radiusof, the cell. Furthermore, knowledge of which
`neighboring cells are to be used in the location determina-
`tion may beusedto further refine selection of such a point
`by possibly selecting a point closest to the neighboringcells.
`An exemplary nominal location estimate 150 is shown in
`FIG. 1.
`
`Given the initial nominal location estimate, and therefore
`the known distances (D,,;) between the nominal location
`estimate and the i’th transceiver,
`the correction values
`included in Ap are calculated as the difference between the
`measured pseudo-ranges (D;) and the nominal pseudo-
`ranges. Thus, for a location determination involving M22
`.
`> .
`sites, A, is defined as:
`
`Do - Dro
`
`D; - Dy
`
`Du-1 — Dai
`
`The direction cosine matrix c includes direction cosines for
`each of the M sites, and is defined as:
`
`7
`
`

`

`5,929,806
`
`Cxn0
`
`Cyno
`
`C=]
`
`Crs
`
`Cyni
`
`CxnM—-1 CynM-1
`
`Denoting the known location of the 1’th neighboringsite as
`(x,y,), the elements of ¢ are defined as follows:
`
`8
`understood that, rather than a single upper threshold, sepa-
`rate thresholds, U,, and U,, having different values may be
`used for the separate divergence conditions. If it is deter-
`mined that divergence has occurred, further iterations of the
`WLSsolution are discontinued, and no location estimate is
`determined.
`If the divergence conditions are not met, it is determined
`at step 512 whether a magnitude of the nominal location
`correction vector has converged below a lowerthreshold, L,
`as described by the equation:
`
`Cxni =
`
`An — Xj
`
`[Ay | = [62 + 02 <L
`
`V Gn P+ On — ye?
`Yn — Yi
`
`V On VP + On — ye?
`
`Cyni =
`
`_
`
`Having determined A , and c, a nominallocation correc-
`>
`tion vector A ,,, can be determined at step 504. The nominal
`location correction vector contains corrections for the nomi-
`nal location estimate and is defined as:
`
`15
`
`20
`
`25
`
`Using linear algebra to solve WLS location solution, the
`nominallocation correction vector is given by the equation:
`
`30
`
`= A
`
`y =(CTVcyCTVAp
`
`The value of L is a matter of design choice, and in simula-
`tions has beenset to a value of 8.05 meters (0.005 miles). If
`the nominal location correction vector has not converged
`below the lower threshold, processing returns to step 504
`where the WLSsolution is allowed to iterate another time.
`However,
`if the nominal
`location correction vector has
`converged,
`the location estimate is set equivalent to the
`current nominal location estimate at step 514.
`The present invention provides a method for estimating a
`location of a mobile unit
`in a wireless communication
`system using a common time base between the mobile unit
`and fixed transceivers. Because a common time base is
`available at both the mobile unit and fixed transceivers,
`signals from only two or more fixed transceivers are
`required. Where the common time base is provided by GPS
`receivers,
`the present
`invention may be beneficially
`employed as an interim location solution for those periods
`where a GPSlocation is not available.
`Although the present invention has been described in
`terms of the presently preferred embodiments, it will be
`readily apparent
`to those skilled in the art
`that various
`modifications can be made therein without departing from
`Where v is a weighting matrix which effectively weights
`the spirit of the present invention or from the scope of the
`each pseudo-range according to its corresponding measure-
`appended claims.
`mentreliability. Although other techniques for computing v
`We claim:
`are known,a preferred technique is disclosed in U:S. patent
`1. In a wireless communication system comprisingafixed
`application Ser. No. 08/792,331 entitled METHOD AND
`APPARATUS FOR COVARIANCE MATRIX ESTIMA-
`infrastructure in wireless communication with a mobile unit,
`&5
`TION IN A WEIGHTED LEAST SQUARES SOLUTION
`the fixed infrastructure further comprising at least two fixed
`transceivers, a method for estimating a location of the
`filed Jan. 31, 1997.
`mobile unit, the method comprising steps of:
`location
`At step 506, having solved for the nominal
`receiving, by the mobile unit, signals transmitted by each
`correction vector, the nominal location estimate is updated
`=>
`of the at
`least
`two fixed transceivers to produce
`by adding A,,,i.c., (x,+5,,y,+5,). With this new nominal
`received signals, each of the received signals including
`location estimate, both the pseudo-range correction vector
`_>
`a timeof transmission indication generated by a respec-
`tive one ofthe at least two fixed transceiversrelative to
`a common time base;
`determining, by the mobile unit, a time of reception
`indication relative to the commontime base for each of
`
`35
`
`45
`
`50
`
`the received signals, wherein the commontimebase is
`provided by a continuously corrected local oscillator;
`for each of the received signals, determining a pseudo-
`range based on the respective time of transmission and
`time of reception indications; and
`determining a weighted least squares solution using the
`pseudo-ranges to estimate the location of the mobile
`unit.
`2. The method of claim 1, further comprising a step of:
`sending, by the mobile unit,
`the time of transmission
`indications and the time of reception indications to a
`location processor forming part of the fixed infrastruc-
`ture;
`wherein the steps of determining the pseudo-ranges and
`determining the weighted least squares solution are
`performed by the location processor.
`
`55
`
`65
`
`8
`
`A p and the direction cosine matrix ¢ are also recalculated
`to reflect the correction to the nominal location estimate.
`.
`.
`=>
`=>
`This process of calculating Ap, c and A,,;
`updating the
`.
`.
`.
`. >
`nominal location estimate; and re-calculating A,, ¢ and
`, can be iterated until a final solution is reached, as
`described below.
`
`> A
`
`Atstep 508, it is determined whether the nominal location
`correction vectoris diverging. This is done by comparing the
`components of the current nominal location correction vec-
`tor Bxcurren» 9y,curren) With the components of the previ-
`ously calculated nominal
`location correction vector (6,,
`previous,Oyprevious: I particular, divergence is deemedto have
`occurred when the following conditions are met:
`
`Sxcurrent?U'S,'y,previous
`oy,
`)yprevious
`8 current?U'6,
`
`U is an upper threshold having a value greater than unity
`and, in the preferred embodiment, has a value of 10. It is
`
`8
`
`

`

`5,929,806
`
`10
`for each of the received signals, determining a pseudo-
`range based on the respective time of transmission and
`time of reception indications; and
`determining, by the location processor, a weighted least
`squares solution using the pseudo-ranges to estimate
`the location of the mobile unit.
`
`10. The method of claim 9, further comprising a step of:
`sending, by the at least
`two transceivers,
`the time of
`transmission indications and the time of reception
`indications to the location processor;
`wherein the step of determining the pseudo-ranges is
`performed by the location processor.
`11. The method of claim 9, further comprising steps of:
`determining, by each of the at least two transceivers, the
`pseudo-rangefor the corresponding oneof the received
`signals; and
`sending, by the at least two transceivers, the pseudo-
`ranges to the location processor.
`12. The method of claim 9, the step of determining the
`weighted least squares solution further comprising sub-steps
`of:
`
`>
`.
`ee
`determining a pseudo-range correction vector A, and a
`direction cosine matrix ¢ based on a nominal location
`estimate;
`
`_
`
`location correction vector A
`calculating a nominal
`according to the equation:
`
`xy
`
`where v is a weighting matrix;
`
`updating the nominal location estimate, the pseudo-range
`correction vector, and the direction cosine matrix based
`on the nominal location correction vector; and
`iterating the previous steps of calculating and updating so
`long as the nominal location correction vector is not
`diverging and so long as the nominal location correc-
`tion vector has not converged beyondat least one lower
`threshold.
`
`13. The method of claim 12, the step of determining the
`weighted least squares solution further comprising a sub-
`step of:
`setting the location of the mobile unit equivalent to the
`nominal location estimate when the nominal location
`
`correction vector has converged beyondthe atleast one
`lower threshold.
`
`14. The method of claim 12, the step of determining the
`weighted least squares solution further comprising a sub-
`step of:
`discontinuing the iteration of the steps of calculating and
`updating when the nominal location correction vector
`has diverged beyond at least one upper threshold.
`15. The method of claim 9, further comprising a step of:
`transmitting, by the location processor, the location of the
`mobile unit to the mobile unit.
`
`10
`
`15
`
`25
`
`30
`
`35
`
`40
`
`45
`
`50
`
`55
`
`60
`
`9
`3. The method of claim 2, further comprising a step of:
`transmitting, by the location processor, the location of the
`mobile unit to the mobile unit.
`4. The method of claim 1, wherein the steps of determin-
`ing the pseudo-ranges and determining the weighted least
`squares solution are performed by the mobile unit.
`5. The method of claim 4, further comprising a step of:
`transmitting, by the mobile unit, the location of the mobile
`unit to an infrastructure entity of the fixed infrastruc-
`ture.
`
`6. The method of claim 1, the step of