US006070079A
`6,070,079
`(114) Patent Number:
`United States Patent 55
`Kuwahara
`[45] Date of Patent:
`May30, 2000
`
`
`[54] POSITIONING APPARATUS USED IN A
`CELLULAR COMMUNICATION SYSTEM
`AND CAPABLE OF CARRYING OUT A
`POSITIONING. WITH A HIGH ACCURACY
`IN URBAN AREA
`
`5,929,752
`5,930,243
`5,930,717
`5,933,114
`
`7/1999 Janky et ab. eee ceeceseeeeseeeeeee 340/988
`7/1999 Parish et al. wee 455/562
`
`
`.. 455/456
`7/1999 Yost et al.
`........
`8/1999 Eizenhofer et al.
`sb eeeeseseeseeeeesees 455/456
`OTHER PUBLICATIONS
`
`[75]
`
`Inventor: Yoshihiko Kuwahara, Tokyo, Japan
`
`[73] Assignee: NEC Corporation, Japan
`
`[21] Appl. No.: 09/234,577
`[22]
`Filed:
`Jan. 21, 1999
`
`[30]
`
`Foreign Application Priority Data
`
`Kikuchi, Hideo et al., Technical Report of IEICE, “Simul-
`taneous Estimation of the Incident Direction and Propaga-
`tion Delay Time of a Multiple Wave Using 2D-Unitary
`ESPRIT,” vol. 97-98, Jul. 1997, pp. 53-60.
`Primary Examiner—Wellington Chin
`Assistant Examiner—Simon Nguyen
`Attorney, Agent, or Firm—Ostrolenk, Faber, Gerb & Soffen,
`LLP
`
`Japan woeceseeeeceeee 10-009908
`[JP]
`Jan. 21,1998
`[51]
`Int. C17eee H04Q 7/20; H04Q 7/36;
`HO4B 7/02
`[52] U.S. Ch. ceceeeeeeee 455/456; 455/457; 455/561;
`455/562; 340/988; 340/991; 340/993; 342/58;
`342/109; 342/450; 342/451; 342/457
`[58] Field of Search occ 455/456, 457,
`455/561, 417, 562, 132; 342/357.01, 450,
`457, 58, 109, 113, 115, 451, 454, 461;
`340/988, 991, 993
`
`[56]
`
`.
`References Cited
`U.S. PATENT DOCUMENTS
`
`ABSTRACT
`[57]
`A positioning apparatusis provided in a cellularbasestation
`for positioning a portable terminal in a cell covered by the
`cellular base station. The positioning apparatus has a base
`position indicative of a position of the cellular base station
`on a map. The positioning apparatus comprises an array
`antenna for receiving a transmission signal transmitted by
`the portable terminal
`to output a plurality of reception
`signals. A receiver section translates the reception signals
`into a plurality of baseband signals to demodulate the
`baseband signals into a plurality of demodulated signals. An
`estimation section estimates incident direction and delay
`time of the transmission signal on the basis of the demodu-
`lated signals to output an estimation result indicative of the
`incident direction and the delay time. A position calculating
`5,819,199
`10/1998 Kawai et al. w.ccececeereeeeee 342/357
`oe oyhoon vt ereserneanencanencencnnsaseasanenetns feea meanscalculates a terminal position of the portable terminal
`
`.
`.
`.
`:
`5873,
`..seeescesscescescencenceseseeseeensens
`UNL
`scccssssssssssssssseeeeesen 342/457
`5,883,598
`3/1999 Parl et ale
`onthemapinaccordancewith the elemek resultandihe
`3/1999 Chang et al. wvvsesssssssessssssseen 455/562
`5,890,067
`postiio
`Pur a terminal Poston signal Inacalve
`
`6/1999 Itoh scesesesssesssssssssssassessesnseeen 340/088
`5,911,774
`‘OF the terminalposition.
`
`
`6/1999 Beasley ...cceccecesceeseseeseneee 455/562
`5,918,154
`7/1999 Lewineret al. wee 455/562
`5,926,768
`
`8 Claims, 8 Drawing Sheets
`
`
`
`PT : PORTABLE
`TERMINAL
`
`101
`¢
`ji...)
`101-1
`
`
`
`103-1
`102-1
`2
`a
`
`
`
`
`CDMA
`
`
`DEMODULATOR
`RECEIVER
`
`PDA
`POWER
`DELAY
`ANGLE
`
`a
`a
`ESTIMATION
`
`
`
`
`CDMA
`CIRCUIT
`
`
`
`RECEIVER
`DEMODULATOR
`
`
`106
`rv
`
`POSITIONING CIRCUIT
`TRANSMITTER
`
`
`
`
`
`104
`
`2
`
`
`c
`
`CELLULAR : SB
`BASE STATION
`
`APPLICATION
`SYSTEM
`
`APPLE 1007
`
`APPLE 1007
`
`1
`
`

`

`U.S. Patent
`
`May30, 2000
`
`Sheet 1 of 8
`
`6,070,079
`
`vOLL-€0LL-ZOL
`
`NOILVWILSS3
`
`LINDYID
`
`Y3AMOd
`
`aTONY)W-E€0LW-ZOL
`SIONV_.
`
`
`Vdd
`
`
`YOLVINGOWSd
`vWadH3SAIO0sgY
`
`1‘SOL901
`
`YOLVNGOWSQ
`vWa)dd35AI303s4
`Z0L——=|O00
`
`/\\TIGVLYOd:id
`
`
`
`TWNIWNYALL
`
`
`
`
`
`LINDYIDDNINOILLISOdYSLLINSNVYL
`
`NOLLYOMddV¥
`
`WALSAS
`
`ASVdStgS-aeWINTONOLLVLS
`
`
`
`LOld
`
`2
`
`
`
`

`

`U.S. Patent
`
`May30, 2000
`
`Sheet 2 of8
`
`6,070,079
`
`
`
`INPUT
`
`> OUTPUT
`
`
`3
`
`

`

`U.S. Patent
`
`May30, 2000
`
`Sheet 3 of 8
`
`6,070,079
`
`START
`
`TO OBTAIN SPATIAL AVERAGE
`
`DIVIDE ACQUIRED DATA INTO
`MXN-DIMENSIDNAL MATRICES
`OF (N—N+1) X (M—M+1)
`AS SHOWNIN FIG.5
`
`TRANSLATE MXxN-DIMENSIONAL
`COMPLEX MATRIX
`OF RECEIVED DATA xnm
`INTO VECTOR TO GENERATE
`CORRELATION MATRIX RS
`
`
`
`
`
`
`
`
`
`S1
`
`$2
`
`$3
`
`S4
`
`$5
`
`36
`
`$7
`
`PERFORM SPATIAL AVERAGE
`
`PERFORM
`UNITARY TRANSLATION ON RS
`
`
`
`
`OBTAIN ENSEMBLE AVERAGE
`OF S1~S4 TO EXTRACT REAL PART
`OF CORRELATION VECTOR
`
`FACTORIZE Ry INTO EIGENVALUES
`
`
`
`
`
`ESTIMATE WAVE NUMBER (L)
`BY ARRANGING EIGENVALUES
`IN DESCENDING ORDER
`
`FROM PARTIAL SIGNAL SUBSPACE Es
`
`
`
`
`GENERATE NEXT MATRIX
`TO FIND INCIDENT DIRECTION
`USING TLS-ESPRIT METHOD
`
`DEFINE NEXT MATRIX Exy
`
`$10
`
`4
`
`

`

`U.S. Patent
`
`May30, 2000
`
`Sheet 4 of8
`
`6,070,079
`
`CALCULATE ExyHExy
`TO FACTORIZE IT
`INTO EIGENVALUES
`
`FACTORIZE MATRIX E
`INTO LXL— DIMENSION
`FOUR MATRICES
`
`S11
`
`$12
`
`CALCULATE MATRIX Vu
`
`S13
`
`OBTAIN Wv SIMILARLY
`TO S9~S13
`
`FACTORIZE MATRIX
`He + +e
`utjVv
`INTO EIGENVALUES
`
`FIND INCOMING DIRECTION
`AND DELAY TIME
`
`314
`
`S15
`
`R MD
`
`OBTAIN SIGNAL POWER
`
`S17
`
`5
`
`

`

`U.S. Patent
`
`May30, 2000
`
`Sheet 5 of 8
`
`6,070,079
`
`AONANDAYSA
`
`AVY
`
`GS‘Jld
`
`6
`
`

`

`6,070,079
`
`9‘JIA
`
`YaMOd
`
`aAlLy138ALVINWO
`
`
`
`YaMOd“IWNOIS
`
`Sheet 6 of 8
`
`
`
`aWILAVTadSSMVANSOIS
`
`OLNI
`
`“mhA
`
`NOILVZIYOLOVA
`
`SSNIVWANSOIF
`
`O.1LNI
`
`U.S. Patent
`
`May30, 2000
`
`ASWLINN +
`
`
`ONIGHODDV«[STWNDISNOLLWISNVU_L
`SSa0ONda]IvitavdWxOS
`
`GOHLAWLlddSa
`SilOLaovdsans
`
`“WILLVdS
`
`ADVYSAV
`
`WYOASNVAL
`
`Nowly
`
`+/\<~W
`
`YalyNOs
`
`
`NOWLOa4ICAAT+ORWHOASNVYL
`XIMLVA
`
`AZINOLOVSualNOS
`
`7
`
`
`
`
`
`
`
`
`
`

`

`U.S. Patent
`
`May30, 2000
`
`Sheet 7 of 8
`
`6,070,079
`
`0zol
`
`0
`
`0Z—
`
`SALTSY
`
`
`
`
`
`canencaannaPawneeanytensghnanalannah-YAMOd
`
`NOILLOSYIG
`
`(3393)
`
`
`
`OeOe—(QP)
`
`09—os—
`
`Or—
`
`ZL‘Sls
`
`AWILAVTad
`
`(9987!)
`
`8
`
`
`

`

`U.S. Patent
`
`May30, 2000
`
`Sheet 8 of 8
`
`6,070,079
`
`VOL
`
`SOL
`
`WALSAS
`
`NOILVYOMddV
`ONINOILISOd
`
`901
`
`NOILVYWILSA
` LINDAID
`
`LINDYID
`
`
`
`Peetennne,
`
`poccoehwees,
`
`NOILLVY¥aivo
`
`TWNDIS
`
`YOLVYSANAD
`
`9
`
`

`

`6,070,079
`
`1
`POSITIONING APPARATUS USED IN A
`CELLULAR COMMUNICATION SYSTEM
`AND CAPABLE OF CARRYING OUT A
`POSITIONING WITH A HIGH ACCURACY
`IN URBAN AREA
`
`BACKGROUND OF THE INVENTION
`
`This invention relates to a positioning apparatus
`(hereinafter a positioner) and more particularly to a posi-
`tioner in a cellular base station that determines a position
`(positions) a portable terminal in a service area (cell) cov-
`ered by the cellular base station.
`Various methods of connection are employed in mobile
`communication systems such as portable telephone systems.
`Specifically, the methods of access can be generally classi-
`fied into three methods,
`i.ec., FDMA (frequency division
`multiple access) method, TDMA (time division multiple
`access) method and CDMA(code division multiple access)
`method. It has already been determined to use the CDMA
`method among those three methods for connection for IMT
`(international mobile telecommunication) 2000 which is the
`next generation portable telephone system because it
`employs a spread spectrum technique and therefore exhibits
`higher efficiency of frequency utilization compared to the
`TDMAmethod.
`
`While there are many types of spread spectrum techniques
`(e.g., the direct spread (DS) method and frequency hopping
`(FH) method), this invention is applied to DS type spread
`spectrum communication (CDMA).
`it is
`In such a CDMAtype portable telephone system,
`desirable to position portable terminals. Such a need arises
`in consumer services and emergency car guiding systems.
`Consumer services are required by the IMT 2000 as an
`alternative to GPS navigation. Emergency call services are
`under evaluation and experiment in the United States as
`emergency car guiding systems.
`Two methods described below have been used as systems
`for positioning portable terminals. One is a positioning
`system utilizing the GPS (global positioning system)
`method.In this positioning system, each portable terminalis
`equipped with a GPSreceiver to allow the portable terminal
`to position itself, and the result of positioning at each
`portable terminalis transmitted to a base station to allow the
`base station to manage the information on the positions of
`the portable terminals.
`The other is a positioning system utilizing a direction
`finder. In this positioning system, a plurality of base stations
`are provided with an azimuth detecting function, and a
`portable terminal is positioned based on a point at which
`azimuth lines defined by azimuth detection signals output by
`the base stations. The result of positioning may be trans-
`mitted from each basestation to a portable terminal to allow
`its portable terminal to identify the position of itself.
`However, the above-described positioning system utiliz-
`ing the GPS method has problems as described below. The
`first problem is that it is subject to limitations during use in
`an urban area. The reasonis that positioning requires recep-
`tion signals from 3 or 4 CPSsatellites, and it is difficult to
`identify locations that are blocked by buildings, roadside
`trees and the like. The second problem is that positioning
`accuracy is low. The reasonis that positioning signals from
`GPSsatellites open to public use are intentionally manipu-
`lated to provide low positioning accuracy in favor of the
`national interest of the United States.
`
`10
`
`15
`
`20
`
`25
`
`30
`
`40
`
`45
`
`50
`
`55
`
`60
`
`65
`
`2
`problem is that it can not be used in existing cell configu-
`rations for portable telephones. The reason is that
`the
`positioning based on the output of a plurality of direction
`finders requires sufficient overlaps between the cells which
`does not currently exist. The second problem is that it is
`subject to limitations during use in an urban area. The reason
`is that it is necessary to maintain line of sight from several
`base stations to a portable terminal which is often difficult in
`an urban area.
`
`SUMMARYOF THE INVENTION
`
`It is therefore an object of this invention to provide a
`positioning apparatus hardly affected in an urban area.
`It is another object of this invention to provide a posi-
`tioning apparatus capable of positioning portable terminals
`without changing cell configurations for portable tele-
`phones.
`It is still another object of this invention to provide a
`positioning apparats capable of carrying out positioning with
`high accuracy.
`On describing the gist of this invention, it is possible to
`understand that a positioning apparatus is provided in a
`cellular base station for positioning a portable terminal in a
`cell covered by the cellular base station. The positioning
`apparatus has a base portion indicative of a position of the
`cellular base station on a map.
`According to this invention, the positioning apparatus
`comprises an array antenna for receiving a transmission
`signal transmitted by the portable terminal, receiver means
`for translating the reception signals into a plurality of
`baseband signals to demodulate the baseband signals into a
`plurality of demodulated signals, estimation meansfor esti-
`mating incident direction and delay time of the transmission
`signal on the basis of the demodulated signals to output an
`estimation result indicative of the incident direction and the
`
`delay time, and position calculating meansfor calculating a
`terminal position of the portable terminal on the map in
`accordance with the estimation result and the base position
`to output a terminal position signal indicative of the terminal
`position.
`BRIEF DESCRIPTION OF THE DRAWINGS
`
`FIG. 1 is a block diagram of a cellular base station having
`a positioner according to a preferred embodiment of this
`invention along with a portable terminal.
`FIGS. 2A to 2C show a principal of generation of a
`synchronoussignal using matched filters provided in receiv-
`ers of the cellular base station illustrated in FIG. 1;
`FIG. 3 is a flow chart of a first half of a procedure for
`estimating incident directions and delay times according to
`the 2D-Unitary ESPRIT method;
`FIG. 4 is a flow chart of a second half of a procedure for
`estimating incident directions and delay times according to
`the 2D-Unitary ESPRIT method;
`FIG. 5 is an illustration showing a method for spatial
`averaging according to the 2D-Unitary ESPRIT method;
`FIG. 6 is a block diagram showing processing steps
`according to the 2D-Unitary ESPRIT method;
`FIG. 7 is a graph showing an exampleofthe result of PDA
`estimation carried out according to the 2D-Unitary ESPRIT
`method; and
`FIG. 8 is a block diagram showing a configuration of a
`positioner illustrated in FIG. 1.
`DESCRIPTION OF THE PREFERRED
`EMBODIMENT
`
`The above-described positioning system utilizing a direc-
`tion finder has problems as described below. The first
`
`Referring to FIG. 1, the illustrated cellular base station SB
`comprises an array antenna 101 havingfirst through M-th
`
`10
`
`10
`
`

`

`6,070,079
`
`3
`antenna elements 101-1 through 101-M for receiving signals
`transmitted by the portable terminal PT, where M represents
`a positive integer which is not less than one. First through
`M-th receivers 102-1 through 102-M are for translating
`signals from the respective antenna elements of the array
`antenna 101 into baseband signals. First
`through M-th
`CDMAdemodulators 103-1 through 103-M are for demodu-
`lating a transmission signal multiplexed using the CMDA
`(code division multiple access) method to output demodu-
`lated signals in a plurality of systems. A PDA (powerdelay
`angle) estimation circuit 104 is for estimating an incident
`direction, delay time, and relative power from the demodu-
`lated signals in a plurality of systems and is for outputting
`the result of estimation indicating the incident direction,
`delay time, and relative power. A positioning circuit 105 is
`for establishing the position of the portable terminal PT on
`the basis of the incident direction and delay time indicated
`by the result of estimation output by the PDA estimation
`circuit 104 and is for outputting the result of position of the
`portable terminal PT. A transmitter 106 is for transmitting
`the result of positioning to the portable terminal PT through
`a transmission antenna 107. The combination of the receiv-
`
`4
`Description will be made on a method for estimating the
`incident direction and delay time of a plurality of correlated
`incident signals with reference to Hideo Kikuchi, Nobuyoshi
`Kyuma and Naoki Inagaki, “Simultaneous Estimation of the
`Incident Direction and Propagation Delay time of a Multiple
`Wave Using 2D-Unitary ESPRIT” Technical Report of
`IEICE, vol. 97-98, July 1997 (hereinafter referred to as
`article).
`“2D-Unitary ESPRIT” is a method which was conceived
`to perform two-dimensional parameter estimation
`(direction/angular height, incident direction/delay time and
`the like) and which has the followings features.
`(A) It requires no knowledge of the response (steering
`vector) of array elements at the antenna aperture.
`(B) It does not require peak search based onthe steering
`vector.
`
`The relationship of rotational invariance in “2D-Unitary
`ESPRIT” is expressed by the following Equations (1) and
`(2).
`
`10
`
`15
`
`1
`Hi
`()
`tan [F[Kuna Vi) = Kyod(ui, vi)
`ers and CDMA demodulators is referred to as receiving
`device. Among the above-described components, the trans-
`V;
`2
`tan [=|v vi) = Ky2d(ui,vi)
`2)
`25
`mitter 106 and transmission antenna 107 maybeatrans-
`mitter and a transmission antenna shared by cellular base
`stations SB.
`
`Reviewing FIG. 1, the portable terminal PT transmits a
`spread spectrum signal which has been subjected to second
`order modulation using a predetermined second order modu-
`lation code. The cellular base station SB receives a multi-
`plexed spread spectrum signal transmitted by a plurality of
`portable terminals PT in a service area (cell) covered by the
`cellular base station SB. As is well-knownin this technical
`
`field, in order to demodulate a spread spectrum signal, it is
`necessary to capture the spread spectrum signal in synchro-
`nism therewith and to demodulate the synchronously cap-
`tured spread spectrum signal using a second order modula-
`tion code assigned to the portable terminal PT.
`Referring to FIGS. 2A to 2C, description will be made as
`regards synchronous capture carried out with a matched
`filter provided at each receiving device. FIG. 2A indicates an
`input signal input to the matchedfilter. FIG. 2B indicates the
`configuration of the matched filter. FIG. 2C indicates an
`output signal output by the matched filter. As shown in FIG.
`2B, the matchedfilter comprises a delay line 108, a plurality
`(seven in the illustrated example) of tape 109, a plurality
`(seven in the illustrated example) of weighting circuits 110,
`and an adder 111.
`
`Time sequence signals with a period T having codes
`corresponding to coefficients set in the seven weighting
`circuits 110 (1, 1, 1, -1, -1, 1, -1 in the example in FIG. 2B)
`as shown FIG. 2Aare input to the matchedfilter. In this case,
`as shownin FIG. 2C,auto-correlation is maximized to show
`a peak at timing whenan input code agrees with a weighting
`code. Each of the CDMA demodulators 103-1 through
`103-M demodulates using such a peak as a synchronous
`pulse.
`Whenthe incident direction and delay time of a portable
`terminal PT is estimated from a plurality of demodulation
`signals received by the array antenna 101, consideration
`must be paid on multi-path signals produced byreflection
`and diffraction due to buildings and the like especially in an
`urban environment. Specifically, it is necessary to select a
`direct wave from among a plurality of incident signals
`having correlation with each other and to find the incident
`direction and delay time of the same.
`
`30
`
`35
`
`40
`
`45
`
`50
`
`55
`
`60
`
`65
`
`11
`
`The definitions for the parameters used here are as fol-
`lows. «1 and vi are respectively expressed by Equations (3)
`and (4).
`
`H=2nfoAd sin 0x
`
`Vv=-20Aft;
`
`3)
`
`(4)
`
`The symbols in the above Equations (3) and (4) are given
`by:
`f0: carrier frequency;
`Ad: element interval;
`61: incident direction of an i-th incident wave;
`Af: frequency sweep interval;
`ti: delay time of an i-th incident wave.
`Kul and Ku2 in Equation (1) are expressed by Equation
`
`6.
`
`Kut = Iu @ Ky
`
`Kia = Iu @ Ky
`
`(6)
`
`where (X) represents a Kronecker operator.
`K1 and K2 in the above Equation (6) are expressed by
`Equation (7).
`
`K, =Re[Qf_ Ow]
`
`Ky = Im[OF_ Ow]
`
`7)
`
`where Re[-] represents a real part and Im[-] represents an
`imaginary part.
`Kv1 and Kv2 in Equation (2) are expressed by Equation
`(8).
`
`Ky; = K3 @ In
`
`(6)
`
`11
`
`

`

`5
`-continued
`
`Kyo = Ky @ Iy
`
`6
`in Equation 15 are respectively
`d,(ui) and dM(vi)
`expressed by Equations (16) and (17).
`
`6,070,079
`
`K3 and K4in the above Equation (8) are expressed by 5
`Equation (9).
`
`dy(Uj=On4an(ui)
`
`dng(V)=Ons"an(V,)
`
`(16)
`
`(17)
`
`Kz; =RelQ4y | 2Qu]
`Ky = Im[Qq,_2Qu]
`
`(9)
`
`aN(ui) in Equation (16) and aM(vi) in Equation (17) are
`respectively expressed by Equations (18) and (19).
`
`10
`
`ay (i) =
`
`expt jN, uit
`
`expt jNy Hi}
`
`
`
`(18)
`
`Equation (18) represents a steering vector of the incident
`direction.
`
`ay Vi) =
`
`expt jMvi}
`
`expt jMy vi}
`
`
`
`(19)
`
`25
`
`Equation (19) represents a steering vector of the delay
`time.
`
`The suffixes M, N, IM,and IN in Equations (6) through
`(9) are given by:
`M: The numberof divisions of a frequency domain;
`N: The number of array elements;
`IM: M-dimensional order unit matrix;
`IN: N-dimensional order unit matrix;
`J2 in Equations (7) and (9) is expressed by Equation(11).
`J2: selection matrix from the second through M(N)-th 20
`rows having the following format.
`
`15
`
`01
`5
`hel:
`
`0
`
`0
`
`0
`
`1
`
`(11)
`
`QM or QNin Equations (7) and (9) represents an unitary *°
`matrix. When M(N) is an even number (M(N)=2K),
`the
`unitary matrix is given by:
`
`ilk |
`[le
`L
`a
`ox V2 Hx —jlg
`
`35
`
`(12a)
`
`When M(N)is an odd number (M(N)=2K+1), the unitary
`matrix is given by:
`
`jill
`90
`Ik
`1
`oF
`OQoes1 =—=| 0? V2
`;
`v2
`Ie
`0 —jllg
`
`
`
`(12b)
`
`40
`
`45
`
`where O, IK, T; and IIK in Equations 12a and 12b are as
`expressed by Equation (13).
`
`1
`
`0
`
`Ug =
`
`0
`
`1
`
`where 0: 0 vector;
`IK: K-dimensional unit matrix;
`T: transpose of matrix.
`in Equations (1) and (2)
`Furthermore, d(gi, vi)
`expressed by Equation (14).
`
`du, ViJ=vec(Duis Vi)
`
`13
`(13)
`
`50
`
`55
`
`is 60
`
`(4)
`
`Mm (12m=M)in Equation (19) and Nn (12n=N) in
`Equation (18) are respectively expressed by Equations (20)
`and (21).
`
`
`M+1
`
`My =m—
`
`Nyon
`
`
`N+1
`2
`
`(20)
`
`21
`(21)
`
`where vec(:) in Equation (14) is given by:
`vec(-)=operator for
`translating an NxM-dimensional
`matrix into an NM-dimensional vector.
`
`FIGS. 3 and 4 show a procedure for estimating incident
`direction and delay time according to “2D-Unitary ESPRIT”
`having the relationship of rotational invariance expressed by
`Equations (1) and (2).
`Acquired data are divided into a matrix as shown in FIG.
`5 in order to perform spatial averaging of the same(step S1).
`The complex matrix of the received data is translated into
`a vector as expressed by Equation (23) to generate a corre-
`lation matrix R* as expressed by Equation (24) (step S2).
`
`Myn=vec(Xim)
`
`Rin hn
`Roa, = Xam Xe
`
`(23)
`
`(24)
`
`This operation is carried out for each subarray. Then, a
`spatial average process is carried out as expressed by
`Equation (25) (step S3).
`
`1
`
`N-N4+1 MA +1
`
`RS.
`
`(25)
`
`D(i, vi) in Equation (14) is expressed by Equation (15). 5
`
`Dus v)-aMedny
`
`Thecorrelation matrix R* is subjected to unitary transla-
`tion as expressed by Equation (26) (step S4).
`
`(15)
`
`12
`
`12
`
`

`

`Qe Onn
`
`Of, - OF @OF
`
`Que ~ Ly @Lg
`
`6,070,079
`
`8
`A matrix Vv is calculated in a mannersimilar to steps S9
`to $13 (step S14). A matrix E is expressed by Equation (36).
`
`(26)
`
`E,= KyKyo Es |
`
`Es
`
`(36)
`
`Subsequently, a matrix (Wu+jPv) is factorized into eigen-
`values as expressed by Equation (37) (step S15).
`
`{BafPJPQA)QT
`
`37)
`
`to (Pu+W¥vj)
`where a term wzu(i)+jwv(i) diagonal
`becomes the eigenvalue (i=1,
`.
`.
`.
`, L). Furthermore, L
`column vectors of T~* are eigen-vectors.
`At step S16, the incident direction is given by:
`
`Hy = InfoAd sind; = 2 tan! w,,(i)
`
`
`Oj = sf=|afoAd
`
`At the step S16, the delay time is given by:
`
`vj = —2nAft, = 2 tan”! wv(i)
`
`tan”! wv(i)
`nAf
`
`T=
`
`(38)
`
`(39)
`
`Finally, the signal power is calculated (step S17). The
`Matrix for the signal power is expressed by Equation (40).
`
`P=T[As-07J,JP"
`
`(40)
`
`where ASis given by Equation (41) and I, is an LxL-
`dimensional unit matrix.
`
`As=diag[AjAn, ... 5 Ay]
`
`(41)
`
`where (X) represents a Kronecker operator.
`Subsequently, the real part of the correlation matrix is
`extracted by performing ensemble averaging of steps S1
`through S4 as expressed by Equation (27) (step S5).
`
`Ry= Re[E[Q4 ROn a]
`
`(27)
`
`Ry is factorized into eigenvalues (step S6) in order to
`obtain results defined by:
`NN)
`Ai: eigenvalue (i=1, . ..
`el: elgen-vector corresponding to Ai
`The eigenvalues are arranged in a descending order to
`estimate a wave number (L) as expressed by Equation (29)
`(step S7).
`
`MEZAcz... ZAp>DAny=... <Aggyeo?
`
`(29)
`
`A signal subspace E, as expressed by Equation (30) is
`formed (step S8).
`
`E=[e, @,---5 &]
`
`(30)
`
`A next matrix E yw as expressed by Equation (31) is
`generated to find the incident direction in accordance with
`the TLS-ESPRIT method(step S9).
`
`Ki Es
`Furl
`Eg
`fe
`Ss
`
`BD
`
`10
`
`bm wn
`
`25
`
`30
`
`35
`
`40
`
`First, a next matrix Eyy as expressed by Equation (32)is
`defined (step 310).
`
`Exye=lKya Es | Kio Es]
`
`(32)
`
`45
`
`Next, E,,” Ex, expressed by Equation (33) is calculated
`to factorize it into eigenvalues (step S11).
`
`Eu Exy =
`
`
`
`(Kya Es)"
`(KipEs)"
`
`
`
`[Ky Es | KwEs] = EAE
`
`(33)
`
`50
`
`where A=diag[/1, 42, ...2L], Ai represents eigenvalues
`(A12,... 2A2L), E=(el |e2 |... Je.,), and ei represents
`eigen-vector correspondingto 1
`The matrix E is factorized into four LxL-dimensional
`matrices as expressed by Equation (34) (step S12).
`
`55
`
`E;
`
`where ASrepresents eigenvalues for a signal.
`A component diagonal to P is the signal power.
`FIG. 6 shows the above-described processing steps
`according to the “2D-Unitary ESPRIT” method in the form
`of a simple block diagram.
`In FIG. 6, each of signals
`received by M antennas is subjected to Fourier transform
`and is supplied to a block for spatial average, unitary
`translation, and factorization into eigenvalues. The block for
`spatial average, unitary translation, and factorization into
`eigenvaluesis a block that performsthe processesat steps S1
`to S6 in FIG. 3. The output of the block for spatial average,
`unitary translation, and factorization into eigenvalues is
`supplied to a block for forming a partial signal subspace
`which performs the processes at steps $7 and S9 in FIG. 3.
`The output of the block for forminga partial signal subspace
`is supplied to a block for processing according to the TLS
`ESPRIT method which performs the processes at steps S9
`through S14 in FIG. 3. Matrixes Yu and Vv outputted from
`(34)
`60
`the block for processing according to the TLS ESPRIT
`En|Ex
`method are supplied to a block for factorizing Pu+jWv into
`eigenvalues, signal power, and matrix calculation where. By
`this block, the processes at steps $15 through S17 in FIG. 3
`are performed to output the incident direction, delay time
`and relative power.
`incident
`FIG. 7 shows examples of the delay time,
`direction, and relative powerof reception signals propagated
`
`
`
`Fe i|E12 |
`
`E;
`
`Then, a matrix Wu as expressed by Equation (35) is
`calculated (step S13).
`
`65
`
`W=-EvlE.x)*
`
`(35)
`
`13
`
`13
`
`

`

`6,070,079
`
`9
`in a multiplex manner from one portable terminal PT and
`arriving at the base station SB according to the “2D-Unitary
`ESPRIT” method. The examples in FIG. 7 are examples of
`five reception signals propagated in a multiplex manner. The
`five reception signals are referred here as first through fifth
`reception signals which have respective increasing delay
`times.
`Thefirst reception signal is a signal with a delay time of
`24 seconds in an incident direction of -60 degrees. The
`second reception signal is a signal with a delay time of 4u
`seconds in an incident direction of -40 degrees. The third
`reception signal is a signal with a delay time of 6s¢ seconds
`in an incident direction of 20 degrees. The fourth reception
`signal is a signal with a delay time of 8 seconds in an
`incident direction of 0 degree. The fifth reception signal is a
`signal with a delay time of 10 seconds in an incident
`direction of 0 degree.
`In the example shown in FIG. 7, the positioning circuit
`105 determines the first reception signal
`in an incident
`direction of -60 degrees which arrivesfirst as a direct wave
`and judges that the distance between the portable terminal
`PT andthe cellular base section SB is 600 meters from the
`
`delay time of 24 seconds. The result of estimation of the
`incident direction and delay time is sent to an application
`system and also transmitted to the portable terminal PT as
`needed.
`Referring to FIG. 8, description will be made as regards
`an example of the positioner. In FIG. 8, Parts having the
`same functions as those in FIG. 1 are indicated by like
`reference numbers. Theillustrated positioner comprises an
`array antenna 101, a PDA estimation circuit 104, a position-
`ing circuit 105, a transmitter 106, a transmission antenna
`107,first through M-th receiving devices 112-1 to 112-M, a
`control circuit 113, a calibration signal generator 114 and a
`calibration signal switch 115.
`The array antenna 101 is formed by first through M-th
`antenna elements 101-1 to 101-M. The array antenna 101
`receives signals transmitted by a plurality of portable ter-
`minals and outputs them to the first through M-th receiving
`devices 112-1 to 112-M throughthe calibration signal switch
`115. Since the first through M-th receiving devices 112-1 to
`112-M havethe same configuration, only the M-th receiving
`device 112-M is illustrated here. The M-th receiving device
`112-M comprises a receiver (RCVR) 102-M,a plurality of
`weightsetting circuits 116, a plurality of matchedfilters 117
`and a multiplexer circuit (MPX) 118.
`The receiver 102-M translates a signal received by the
`M-th antenna element 101-M into a base band signal and
`phase detects to output an IQ-separated signal
`to the
`matchedfilters 117. The matched filters 117 and the weight
`setting circuit 116 are prepared in quantities corresponding
`to the expected maximum number of portable terminals.
`Weights associated with a demodulation code for each
`portable terminal are set by the control circuit 113 in the
`weight setting circuits 116, and the IQ-separated signal is
`demodulated by the matchedfilters 117. Demodulated out-
`puts corresponding to the numberof portable terminals are
`multiplexed in time domain at the multiplexer circuit 118
`and transmitted to the PDA estimation circuit 104. That is,
`the combination of the plurality of weight setting circuits
`116, the plurality of matched filters 117 and the multiplexer
`circuit 118 corresponds to one CDMA demodulator shown
`in FIG. 1.
`The control circuit 113 sends the information of weights
`to the weightsetting circuit 116 depending on the numberof
`access from the portable terminals and controls the multi-
`plexer circuit 118 to cause it to output demodulated signals
`in respective systems simultaneously at preset timing.
`
`10
`
`15
`
`20
`
`25
`
`30
`
`35
`
`40
`
`45
`
`50
`
`55
`
`60
`
`65
`
`10
`The FDA estimation circuit 104 uses the above-described
`“2D-Unitary ESPRIT” method to estimate the delay time
`and incident direction of each portable terminal and trans-
`mits the result of estimation to the positioning circuit 105.
`The positioning circuit 105 finds the positions of the por-
`table terminals relative to the base station on the basis of the
`information indicative of the delay time and incidentdirec-
`tion of each portable terminal. The positioning circuit 105
`supplies the application system and the transmitter 106
`relative position information indicating such relative posi-
`tions along with the information of the absolute position of
`the base station. The transmitter 106 multiplexes the infor-
`mation on the position of each portable terminal according
`to the CDMA method and transmits it
`to each portable
`terminal through an aerial by transmitting antenna 107.
`The calibration signal generator 114 generates a calibra-
`tion signal for calibrating the channel gain between the
`receiving systems and errors in the transmission phase
`throughout the frequency band. When calibration is carried
`out,
`the setting of the calibration signal switch 115 is
`changed in order to supply the calibration signal to thefirst
`through M-th receiving devices 112-1 to 112-M.
`The embodiment having such a configuration makesit
`possible to position a plurality of portable terminals (mobile
`terminals) in a cell with a device in only one base station
`covering the cell. The result of positioning may be displayed
`on the mobile terminals to provide services that replace the
`existing GPS. The result of positioning may be transmitted
`to service systemsat relevant organizations to allow various
`services, e.g., guidance of emergency vehicles.
`Above-cited article A show that the accuracy of incident
`direction estimation is 0.5 degrees and the accuracy of delay
`time estimation is 0.1 nanosecond when there are two
`
`incident wavesin directions different from each other by 25
`degrees with a difference of 5 nanoseconds in delay time.
`For currentcells for portable telephones which havea radius
`of about 1 Km,the accuracy of positioning is dominated by
`direction and is about 9 meters which is a preferable value
`comparedto the accuracyof the positioning 20 to 200 meters
`of GPS.
`
`As readily understood from the above-mentioned
`description,
`the direct wave and the multipath wave are
`separately identified so that the portable terminal is posi-
`tioned on the base of the incident direction and delay time
`of the direct wave. Therefore, it is possible to prevent the
`position service of the portable terminal from limitation in
`an urban area.
`Since the cellular base station estimates not only incident
`directions but also delay times, it is possible to use a cell
`configuration for portable telephones.
`Furthermore,
`it
`is possible to provide the positioning
`service with a high accuracy inasmuch as no intentional
`manipulation is made to reduce accuracy in favor of the
`national interest of one country unlike the case of GPS.
`While this invention has thus far been described in
`
`conjunction with the preferred embodimentthereof, it will
`readily be possible for those skilled in the art to put this
`invention into practice in various other manners.
`Whatis claimedis:
`
`1. A positioning apparatus provided in a cellular base
`station for determining a position of a portable terminal in a
`cell covered by said cellular base station, said positioning
`apparatus having a base position indicative of a position of
`said cellular base station on a map, the positioning apparatus
`comprising:
`an array antenna for receiving a transmission signal
`transmitted by said portable terminal to output a plu-
`rality of reception signals;
`
`14
`
`14
`
`

`

`6,070,079
`
`10
`
`15
`
`12
`11
`second meansfor calculating said portable position on the
`receiver meansfor translating said reception signals into
`a plurality of baseband signals to demodulate said
`basis of said direct wave and the said base position.
`baseband signals into a plurality of demodulated sig-
`5. A positioning apparatus as cl