GB2340353A - Method and device for demapping digital video data - Google Patents

Description

2340353 NETHOD AND DEVICE FOR DEMAPPING DIGITAL VIDEO BACKGRQUISM OF nffi

2MESM-0-N Field of-tbe Inventign The present invention relates to a method and device for dernapping digital video, and more pardculady, to a method and device in an orthogonal frequency division multiplexing system using multiple carrier for dividing regions of received data and providing Chmel State hiformation used in dernappiag.

BackgLo=d of the Mated Art In digital TV transmission sTteras, There are single carrier modulation systems which -uses a single carrier and rulti-carrier modulation systems wbich uses multiple catriers.

Particularly, the digital TV tr=smissiou system can be sorted into a Vestigial Side Band (VSB) system using a single carrier and an orthog'oral. frequency division multiplexing system (OFDK system using multi-catriers.

The OFD.M systm using multi-carrier allows an easy restoration of damaged signals caused by a multiple path cbz=el and, unlike the single cauier, facilitates Single Frequency Is Network (SM. llav a small bandvvridth, each carrier may easily be damaged from channel variations. However, the signal &mage can be adequately restored because in eft the multiple interference charmel only reduces an amplitude of each ammuittiuS channels in the =jitiple-carrier.

The amplitude size of each carrier is generally caLl ed a Channel State Thformation. (CSI), by which the OFDM can obtain idormation indicating the reliability of a received signal. For a Digital Video Broadcastiag-Terzestrial (DVB-7), determined as a European digital TV standard, pilots of a PN sequence are ftw_%rnitted from a ftwLswaission terminal by inserting the PN sequerxe between data to be trans:rnitted. The channel state may be obtained by comparing 1 the pilots of The PN sequence to the original pilots at the time of transmission, because the original pilots are known at a reception tenninal. Accordingly, the pilots may also be used as the CSI, After mapping to a constellation, a data is inverse fast fourier transformed (IFM prior to transmission according to the OFDM. Also, guard intervals are inserted in the data according to a modulation method prior to the transmission. The modulad= method employed az the tran mission side is typically a Quadrature Phase Shift Keymg (QPK), 16- Quadrature Amplitude Modulation (QAM), or 64-QA.M. Thus, to trazzmit. a data according io the OFDM, the data must be mapped in one ofte tbiee modulation method, subjected to EFFT, and inserted of guard intervals therebetween, before transmissim Fig, I is a block diagram of a DVB-T reception system.

The tuner 101 at the receiver side receives an OFDM signal through an ante=a and converts the signal into an Intermediate Frequency (IF) using a co=ol signal ouq= by an Auto Chda C ontrol (AGC) unit 10 6, T"he IF signal from the tuner 10 1 is forwarded and digitized al an A/D converting unit 102, and forwarded to the IIQ generating unit 103. Because the digitized signal from the A/D converting unit 102 only has an iaphase component, the I/Q generating unit converts the digitized signal into a complex sipal baviug a quadrature componm and outputs to a frequency corzectizig unit 104.

In the OFDM reception system, a frequency error occurs between the transmitter and rec6ver, such as an error caused by Iccal oscillation of the tuner 101. Ile error causes a frequency offset which is corrected by a signal known as an Auto Frequency control (AFC) signal. To oorrect the frequency offset, an AFC unit I 10 exuwts information of the frequency offset, using a pilot signal extracted by a pilot extracting unit 108, and outputs the extracted 2 information to the frequency correcting unit 104. The frequency correcting unit 104 multiplies the extracted eormation to the output of the JIQ generating unit 103, to correct the frequency offset, and fcrwards the corrected signal to a Fast Fo=ier Transform TM unit 105, to d3a AGC unit 106, and a timing synchronizing imit 107.

The FFT unit 105 sub ects the output of the frequency correcfmg unit 104 to M with respect to a starting point provided by the trning ry=luonizing unit 107, and outputs the sig7w to the pilot extracdng unit 108 and an equalizing unit 109. Th AGC unit 106 generates a signal to maintain an amplitude of a signal provided to the A/D cmverting unit 102 at an appiroprize level, and controls the tuner 101. The pilot extracting tmit 108 exhwts a scauered pilot signal inserted at the transtnitter side from the received signal, end outputs the octracted pilot isna to the equalizzing unit 109 and to the AFC unit I 10. Mie equEzing unit 109 compensate the carrier which was distorted by the chamel uti the received signal and the emacted pilot Signal, and oueputs the compensated sipal to a demapper 1 12. The demapper 112 demaps the equalized data by a reverse process to the mapping of the received dam. The dmapped data is otqxa to a viterbi decoder 114 through an int dointerleaver 113 to gain a symbol deinterleaver and a bit deinterleaver to be demodulated.

The equalizring unit 109 bas a channel predictor 109-1 and a di-vider 1092. MM Chann I predictor 109-1 compares the extracted pilot si to a known pilot valueand intexpohaes the compared result with respect ro a time axis and afrequency wds +,o generate a channel im respo=e, from,%tich a channel can be predicted. The diMer 109-2 divides the FFT signal by the channei impulse response to compemaze the carriei wbich may be distorted by the channel.

Particularly, the channel pre6ctor 109-1 cawes the extracted scattered pilot signal to be divided by a pilot reference value, which is a signal identical to an original pilot sigwd inserted at the 3 aansmitter side, and extwts a sampled frequency characteristic of the channel. Thereafter, the channel predictor interpolates the mmizted sampledfrequency characterisuc, with respect to the time axis and the frequency axis to predict E-equency charuteristics ft the, =tire frequency cArriers. By interpolation using a known scattered pilot, channel characteristics of am active carrier c= be inferred.

Fig. 2 illustrates an OFDM frame structure of a DVB-T, showing an inserted state oft1w scattemd pilots. The black portions illustrate positions of the scattered pilots, and the whize portions illustrate the active caniers, i.e. the daft to be tin-naitted. The scartmed pilots inserted at the, transmitter side are positioned at every 12 carrien on the frequency axis and at every 4 to symbols on the time axis. The channel predictor 109-1 f1m interpolates along the time axis before interpolating along the fxequenc axis. Since the pilot siguls are present at every 3 carriers on the tirne axis rather at every 12 carrier, interpolation with respect to the time axis educes the interval ofthe pilot carriers along the frequency axis by 1/3.

The divider 109-2 causes the FFT signal, delayed by an amount of time takcn to cxtract and interpolate the scattered pilot signal, to be divided by an output of the channel predictor 109 1 for compensating a distortion by the channel. The oomp=sated signal is output to the demapper 112. The derriapper 112 aWopri ately divides regions ofthe data from the divider 109 2 according to a transmitzed constellation.

Fig. 3 illustrates a region dividing method disclosed in USPAT 5,134,635. Fig. 3 illustrates a region db-iding method at a receiver when data is transmitted in 16-QAM, in which 4 bits form one symbol. Fig. 3 shows an example, where the tra=mission bits along an inphase component axis and a quadrature component axis are mapped to 10, 11, 01, 00 along -33, -1, 1, 3 tmm=ission, as one symbol has 4 biM a Namely, as one symbol hw 4 bits in the 16-QAMI 4 symbol can be divided into 2 bits in the inphase component axis and two bits in the quadrature component axis. Tberefore, a fust output bit b(n) on the inphase component axis and on the Taadratu:re componew a)ds, respectively, can be expressed by Equation (1), below, b(I) or b(3) -1, if Z(n) > 1, z(n), if I > z(n) z 0, z(h), if 0 < z(n) < - 1, and 1, othawise. - - In Equation 1, a dammination is mde whether an output signal z0a, k), representing a 00th carrier in an (n)th symbol, of the divider 109-2 is greater than 1 or less thm -1. Ifthe output signal z(n, k) is greater than 1, the first bits ofthe region are determined to bave been transmitted as 0 and a value of -I is output. If the output si" z(n, k) is lea than - 1, the first bits of the region are determined to have been twsmitted as 1, &ad a value of I is output If the output signal z(n, k) is between I and -1, an opposite value of the output si is output.

Also, a second output bit on the inphase component axis and the quadrature component is axis can be expressed as Equations (2), below.

b(2) or b(4) -1, if z(n) > 3, 2- 1 z(n)', if j z(n) I a 1, and 1, otherwise.

Region division values output can thus be calculated in bit metric according to the fbIlowing Equation (3), where y is a received data and b(k) is a region diNided value, i.e., one of z(n), -z(n), or 2- 1 z(n) 1.

m(k, 1.) - I y-b(k) 12 CSI, if y =1, and m(k,O) - I y-b(k) 12 CSL if y I When the CSIs, representing channel state information of each received data symbols, are extracted at the cbannel predictor 109- 1, aregiou dividing method changes as shown in F-quation (3). The bit metric calculated by Equation (3) conducts a hard decision in which sizes are compared to one another and, I and 0 are determined fbr each bit.

However, in the region division of the first bit on the inphasa component axis and the quadrature component axis of the ieceived data, the region division method described above Borcibly divides an output to a value of -I or I if the received- data exceeds L -I.Al ou th OT th Sb C level of noise may be different in each output, the such dit"emnce has not been taken into consideration, Accordingly, an exact regioz division cannot be made. Likewise, a problem adws for the secozLd bit on the inphase component axis and the quadratare component axis of the received data bepause an output is forcibly set to -1 or I if the received data exceeds 3 or -3, or falls between I and -1.

Al.so, the reliability deteriorates during the bit metric calculation because of the hard decision in wbich the region division value is multiplied by the CSI provided from the chaunel.

predictor 109-1 to determine the value of a bit as I if the multiplied value is greater than a pmet value and as 0 odmrwise. 11ae CSI value is simply multipJ!ed to the region division value, even if the CSI value is too low. In such case, the multiplied value would also be too low, resulting in a low reliability of the quantized 1 or 0.

For the calculation and implemmiatioa at the channel predictor 109-1, the CSI is generally obtained only from a signal power, not from a sipal to noise ratio (SNIR) at each carrier position. Although Us may provide good performance on additive white Gaussian noise (AWGN), it causes an error in a frequency selective noise or an interference because the level of noise has not been taken into amount. Therefore, in order to compensate for the level of noise, 6 a method exists in which a noise power is obtained separately for each caiTier, and an ShR is calculated for use as a CSL However, even this method has an unclear process of a noise prediction, and the hardware to implement the process in which the predicted noise is divided by a signal power to obtain an SNR is very complicated..

SlaQdAn OF THE-2=QNL_ Accordingly, it would be desirable to solve at least the problems and disadvantages of the relwed art.

It would also be desirable to provide a method and device for receiving a digital video, in wbieb the received data is uot forcibly defined to a value at a region greater thm a prow region.

It would also be desirable to provide a method and device for rcceiviag a &giW video, in which a region division value is multiplied to a CS1, and a soft decision is made to preset sectons, It would also be desirable to provide a method and dervice for receiving a digital video, in whicb a value greater dian a preset value is forcibly provided if a CS1 value is received smaller than a preset value.

It would be desirable to provide a method and device for receivi;ug a digital video, in which an Mean Square E-.Tor (-VS-E) is calculated by a time axis averaging proms at an equalizing unit for generating a CSL Additional advantages, of the invention will be set forth in part in the description which follows and in part will become apparent to those having ordinazy SIWI in the art upon examiriation, of the following or may be learned from practice of the invention.

7 Advantage3of the invention may be realized and artained asparticularly pointed out in the appended claims.

Accordingly, the present invention provides a method for receiving a digital video. including the steps of (1) linearly dividing regions of data values on an inphase component wds and a quadrature component wds respectively, and mapping ft data to a constelMon before umnmission according to a moduMon method, (2) calculafing a MSE:kOM the received data by a time aids averagm& to generate chanaeI s= informaton, and (3) multiplying the channel s= information to &e region division value fm quantizing.

Step (2) may further include the squaring a difference between the received data and pilot izformation at a r=eived pilot position and time wds averaging, to calculate an MSE at eacia frequency carder, Mking an inverse number of the MSE to calculate an average in a frequency Tegion; calculating on overall M-- and normalizing and lineaz interpolating the normalized data for generating a CSI for every carrier position.

Step (3) may further include the setting the CSI to a preset value or a value greater the CSI if the CSI is smaller than a preset value, before the multiplication to e region division value Step (3) may also include making a soft decision of the C SI to a preset value if the C SI is smaller d= a preset value, befom the multiplication to the region division val= A device for receiving a digital video embodying the invention may include an equalizb:xg unh adapted to cause a sigaal which -was subjected to fm Fourier transformation to be divided by a channel impulse response detected using a received pilot si gaal for compensating a carrier distorted by a channcl; a region detectimg unit linearly dividing regions of data values on an inphase component axis and a quadrauire component axis, respacdvely, received from the 8 equAlizing unit for reflecting a diftence, of diSt8nCe3; 3 channel StWx infOrnmton generating unit calculating an MSF- from the data fiom the equalizing unit by a time Lxis averaging to genenue channel state information; and a decision and quantizing unit multiplying the cba=Cj State Wfornw4on from the cha=el state information generating unit to a value divided by each region at the region detecting unit and quantizing a diffe=ce of multiplied values.

BRIEF MSCRRMON OF THE MAWINGS _ Embodiments of the invention will be described in detail with reference to the following drawings in which like reference numerals refer to like elements wherein:

Fig. I is a block diagrarn of a DVB-T receiver in the related aM Fig. 2 is a fiwu structure of a DVB-T standard in the related art, showing a transmisdon state of pilot signal inserted in active carriers; Fig. 3 is a reSion dividing method in a 16-QAM, in the related &M Fig. 4 is a block diagram -of a DVB-T receiver in accordance with a preferred enibodirrient of the present invention; Fig. 5 is a block diagram of the CSI generating unit in Fig. 4; Fig. 6 is a graph showing a data region dividing method of ihe embodiment on the first bits on the respective coordinates when mapped in 16-QA-M at a receiver terminal; Fig. 7 is a graph showing a daza region dividing method of the embodiment on the second bits on the respective coordinates when mapped in 16-QA.IA at a re ceiver termirk-a; Fig. 8 is a graph showing an exAmple of a raet-hod for providing channel state information in Fig. 4; Fig. 9 is a graph showing a data region dividing method of the embodiment on the 9 third bits on the respective coordinates when mapped in 64-QAM at a receiver tern2i.W; Fig. 10 is a graph comparing the case wben a region dividing value and a CST value an obtained and quautized in 3 bits according to the embodiment - to a case in the related art; and Fig, I I is a graph compiuing a case when a region dividing value and a CSI value are obtained and quanti2ed in 4 bits according to the embodiment to a cast in the related at.

DEjALBP-DESCR=-N OF IM PREEERRED aMQM= Reference wiU now be made in detail to the prefared embodiments of the preseat invention, examplesof which a:e illustrated in the accompanying drawings. Grener&Uy, in the present method and device for receiving a digital video, the received data is not forcibly defiaed for a region over a preset value during a region division for dcmapping, but is defted linearly to make amore accurate region division. Also, when the CSI is too low, eitl= a preset value is provided or a soft decision is forcibly m-.de to prevent a potential errar caused by a low value of a CST value. Finally, in the prediction of a CSL Le. an SN.R at eact. frequency carrier position, a sigoal power and a noise power are not separately predicted, but an MSE is used in an SNIR is prediction to reduce a control process and to improve a performance under a frequency selective noise enviTonment as compared to a case where only the iVW power is used Fig. 4 is a partial block diagram of the device for receiving a digital. video, i.e. a DVB-T receiver, in accordance with the prefer:red embodiment of the present invention, showing parts reWed to a region detection and CSI generation for demapping. Referring to Fig. 4,:he present device for receiving a digital video includes a region detecting unit 401 dividing regions of bits on an inphase component axis and quadrature component wds of the data output by an equalizer; a CST generating unit 402 generating and outputting a CSI signal from the equalized data; decision and quantizing unit 403 applying the CSI to an outpid of the region detecting unit 401 in order to make a soft decision and quantization; an internal deinterleaver 404 conducting both a symbol deintcrleavmg and a bit interleaving of the quantized data; and a viterbi decoder 405 deterrnining values of I or 0 for the internal deinterleaved datL Fig. 5 shows a block diagrasn of the CSI generating unit in Fig. 4.

Referring to Fig. 5, the CSI generating unit includes a squaring unit 501 squaring a.

difference between the equalized data and The pilot informgion of the scaUered pilot; Va MSE calculating ucit 502 averaging outp= of the squaring unit 501 in a time axis for calculatag an NMSE of each ftequency cai7ier; an inverting unit 503 obtaining an invemed value of the output from the MSE calculafmg unit 502; a nomnIdng unit 504 averaging the inversed outputs of the inverting uWt $03 in a frequency region to' calculate an overall SNT,, and normalizing the and an interpolating unit 505 interpolating the outputs of +Jae normalizing unit 504 to predict a CSI at each carrier position.

In the present device for recziving a digital image, each canier value from the equaLzing is unit is provided to the region detecting =it 40 1. The region detecting tmit 401 outputs the equalized data as a value of an Wropriate reeon aomrding to each bit, whereit the value corresponds also to the transmitted constellation. Tkereafter, each region N-alue is calculated according to a uniform equation to be utiliud as a state information of a bit, i.e. whether the bit is I or 0.

Fig. 6 is a &raph showing a data ivgion dinriding method ofthe r4on detecting unit 401 for the ffiTt bits. 11g. 6 is an example of mapping data in 16-QAM at:& transmitm side, where a 4-bit data forms a QAM symbol, in which a first and a third bits am assigned to an inphase component axis, and a second and a fourth bits am assigned to a quadrature componeut axis.

n Because the bits 10, 116, 01, 11 On the iaPh&se component axis and quadrature,conTonent axis, respectively, are =pped as -3, -1, 1, 3 before transmission, the rtgion detecting unit 401 fim divides data regions corresponding to the first bits on the respective coordinates. Fox example, during the mapping at the transmitter side, if a value of I is transmitted for the negative data mapping and 0 is trarismitted for the positive data mappui& with respect to the y axis or x - 0 as the border line, shown in Fig. 6, a relation between the input and output can be expressed with Equation 4 below.

y M -X (4) Accordingly, when the input is greater than 111, the input is not forcibly limited, but is defined linearly. Tlius, the input can be defined more effectively than the related art in which the ii is forcibly defted as 1 or -1 if the M''Put is greater than 11; (shown in dotted line in Fig. 6).

Fig. 7 is a graph showing a data region dividing method of the present invention for the second bits on the inphase component axis and the quadrature component axis. In the division ofthe second bits, the transmitted bits 1 and 0 may be divided at the values of -2 and 2 as shown in Fig. 7. Thus, the input/output reMon can be expressed by Equation (5) below when the input x is positive and by FAILmtloa (6) when the input x is negative, %I=e the output y is the region value.

y = -X - 2 (5) v.- x+2 (6) When the imput x is smaller than I I I or greater than 13 1, the output is not forcibly defined as I or - 1, but is linearly defined. AcCordingly, the received data is converted at the region detecting unit 401 and output to the decision and quanti unit 403.

While the region detecting unit 401 is dividing the regions, the CSI genmting unit 402 CP IL2 calculates an MSE from the outputs of the equalizing unit by a time a= aven&& process to generate a CSI and forwuds the generated CSI to the decision and quantizing unit 403.

Particularly, a receiver for a terrestrial broadcasting sysE=, such as DVB-T, usually does not have sbarp change of a chano I enviro=ent. As a result, an as=ption that a significaw error will not be gtrierated even if the averaging is made with respect to the time axis to predict an SNR is made.

To predict an SNR at each c4rrier position, the embodiment. calqulates anMSE from the outputs of the divider 109-2 mter than the cbannel predictor 109-1 unit by a time axis averaging process. As the p2ot extractor 108 extracts the scattered pilot signal inswed at the ummnitter side, the squaring unit 5 01 receives the scattered pilot siVW as a refaence value and receives the output of the divider 109-2. The squafing unit 5 01 squares a difference of the two signals, and forwards the resulu to the MSE calculating unit 502. The MSE calcukting unit 502 perlbr= the time axis averaging to calculate the MSE at a sc.-A-tered pilot position, which can be expressed by Equation (7), below.

MSE(ej (IN - e, I') )2) (pkhk + n hk I ilk h,t' hk In Equation (7), k denotes a frequency carrier index, e denotes an output of a (k)th carrier in the divider of the equalizing =it, and pk denotes a reference value. Also, bt denotes an hnpulse response of a channel, h', denotes a predicted value of the hi, nk denotes a noise component, and 0 denotes atime axis avemging process. ALI the above variablos am complex 13 values. Equation (7) shows that the MSE aMoacbgs to a rado of a noise- power to a signal power, i.e., an inverse number of the SNR: whenk'kapproaches to hk. Therefore, the MSE prediction from the outputs of the divider 109-2 shows & result close to ihe SNR prediction- In addition, since the referc:nce value pk only has an inphase component, under an assumption ale an inphaw component and a quadrature component of the noise component ut has the same averages and scatters, a mult of Equation (7) may be simplified with Equation (9), below.

MSE(ek) - (0=9[00) rhe imgfe,])2 in Equation (8) denotes a quadrature component of ek. As seem fmm Austion (8), the MSE calculation is simplified in which the calculation process is, not only shorten4 but also requires no reference value, Thereafter, the inverting unit 503 takes an inverse numbcT of a result from the MSt c4lculating unit 502 and forwards the inverted Nralw to the nommUzing unit 504. Because the NSE is a value calculated at a carrier position (113 of entire carriers in the DVB-I) to which the reftence value is applied, an SNR (i.e. a CSD can be predicted at the position by taking an inverse of the MM Is The izimg unit 504 obtains an average in a frequency region from S.NRs cf all pilot positions, stores the average as an overall SNII, and divides the MI from the inverting unit 503 by the overall SNR to normalize. However, the CSI from The normalizing unit 504 is only from a pilot position, and a CSI at a data position must. be obtaiued. Thus, the interpolating unit 505 subjects the result of the normalizing unit 504 to (O)th or first interpolation to obtain a CSI at the overaU carrier position, and outpms the CSI to the decision and quantizing unh 403. Theoverall SNR from the normali unit $04 represents an average SNR at a rw end of the equalizing unit 109, and =y be applied to other parts (a gAin adjustment in a Loop filter, =h as AFC, adjus=ent of an AGC adaptive speed and etc.) to further improve a receiver performance.

14 ne decision and quantizing unit 403 multiplies the output of the re n detecting unit 910 401 to the CSI from the CSI generating unit 402, and quantizes a diffuence of the multiplied values. As the difference inces, such quantization of the difterence allows a more accurate mapping to either I or 0. Particularly, the value obtained by multiplying the CSI and the region division value is quantized in 8 sections or 16 sections. During Us multiplication, if the CSI is low =h t1w the value falls below a predetermined value c, shown in Fig. 8, a de&ult value or a value d greater than the original CSI is output for the multiplication with the region division value. Also, if the CS] is low, the cha=el state may be poor. Thus, the CSI may f6reiblybe set -to 3 or 4 by a soft decision.

The value of CSI is adjusted bec=e a multiplication of a substantially large output of the region detecling unit 401 to a low CSI results in a low multiplied value, thereby b i quantiz.ed into a low value with a low zellability of 0 or I - On the other hand, the larger the multiplied value, the values becomes closer to 0 or 1. Therefore, ifa weghted value is quantized to 0 or 1, a more effective viterbi decoding can be performed using each CSI information.

Namely, an ouW ofthe decision and quantizing unit 403 is forwarded to the viterbi decoder 405 through the intemal deinterleaver 404 havkg a symbol deinterleaver and a bit deinterleavcr and demodulated Fig. 7 is a graph showing a data region diiriding method of the embodiment for the third bits on rMective coordinates when mapped in the 64-QA-X4 at a receiver terminal. In the tr=swission of data ia 64-"M, 6 bits form one symbol, and if the 6 bits are in an order of y,,.

yl, y:, y3, y. y5, the data on tke inphase component axis an ya, y2, y, and the data on the quadrature component axis are y,, y3. y3. Similex to the cast of 16-QAM., region divisions Qfthe fim and second data on the inphase component and the quadrature component for the 64-QAM - are carried in the manner described with ref=ncc to Figs. 6 and 7. The first bits on the inphase component axis and the quadrature component axis are provided according to Equation (4).

However, becanse the data am expanded from -7 to 7 at the mapping iermin4 the second bits are expressed by Equation (9) below.

Y-- IxI -4 (9) The third bits on the inphase component axis and the quadrature componew wds, shown in Fig. 9, area little more complicated, but can be exprcs4 by.E =pn (10) below whentlic qL input x is smaller than 14 i and by Equation (11) when x is greater than 141 y- jxj -2 (10) YW1XI +6 As in the 16-QAK the output of the region detecting unit 401 is not forcibly capped a-t -I or 1, but defted in a way to mfiect, a difference of distimces as is. In the meandm, if the modulation at the transmitter terminal is QPSK and - 1, 0 on the inphase component axis and thequadrature component wds am mapped to - I and 1, the region division values on respective axes - of the -cceived data can be mc7essed according to Equation (4); Le. if the received datals x, -x is Provided.

Figs. 10 and 11 are STaphs showing results of simulatioas amording to the aforementioned methods, wherein.Fig. 10 illustrates a case when 3 bit quantization is conducted at the decision and quantizing unit 403, and Fig. I I illustrates a case when 4 bit quantization is conducted at the decision and quantizing unit 403, as compared to the related art. In Equations and 11, the plot with triangles represent the related art the plot with circles represent the presem invention.

For example in Fig. 10. ia comparison to the related art the pre sent invention can obtain 16 a CNR pin of approximately IdB at 210 which is gfMCTally a Quasi Error Free (QEF) refererim Thus, the data can be dernoduLated even if the data is transtnittcd at a low power of approximately 1 dB from the tmsmiuer terminal, and the data can be demodulated even if a reception power is as low as approximately IdB.

In s=, the pmerit invention is applicable, not only to TV, but also to all digital communicaion system that employs an OEM Moreover, the presem invention is applicable both to the single carrier method and to the multiple ca=igz method if C. SIs are known. As explained above, the mettod and device according to the present invevdon have the following advantages.

FiM an exact region division can be made during the demapping of a received data by not forcibly capping the datA but only dividmg appropriately so as to reflect a diffc=ce of distancesasis. Second, areliability cau be improved as a region division value fmm the above method is multiplied to a CSI, and -abjected to soft decision and quantization. Third, die aror coming from the mulliplicaiion of the region division value and a too low CSI value can be is prevented by providiag a preset value as the C 91 value or by forcibly maldn soft decision when the CSI is too small. Fourth, reliable CSIs at all carder positions can be genemted with a simple hardware and control as the CSI is generated by calculating an MSE from a signal from the equalizing unit, obtainiug an inverse value of the NISE, and making a Uncar irmapoiation.

The foregoing embodiments are merely exemplary and are not to be wastrued as limiting the present invention. ne, present teachings can be readily applied to other types of apparatuses.

The description of the embodiments is intended to be illustmtive, and not to limit the scope of tbLe claims. -Many alternatives, modifications, and variations will be apparent to those sIdIled in the art.

17

Claims (22)

1. A method for dernapping a received digital video data, mapped to a coustellatiou befbre ftwismission acc-ording to a modulation method, comprising:

(a) linearly dividing regions of received data values on an inphase component axis and a quadratre componett w respectively, and outputting a region division value; (b) calculating a Mean Square Error NSE) from the received data to generate a channel sM informadon (CSI); and (c) multiplymg the generated CSI value to the region division value and outputting a pultiplied value to be quantized.

2. A method. of nlaim 1, wherein in step (a), the region division value of data corresponding to first bits of the received data is provided as -x for the received data X, if the modulation tpe of a transmitter terminal is 16-QAM. and ifbits of 10, 11, 01, 11 on the inphase component mds and the quadrature component axis are respecti,%Tly mapped to -3, -1, 1, 3.

3. A method. o,4,' claim 1, wherein in step (a), the region division value of data is corresponding to second bits of the received data on the iuphase component axis and the quadrature component axis is provided as -I x 1 -2 for the received data x, if the modulation t)r of a transmitter terminal is 16-QAM and if bits of 10. 11, 0 1, 11 on the inphase component axis and the quadrature component axis are respectively mapped to -3, -1, 1, 1

4. A method of claL-n, 1, wherein in step (a), the region division value of data corresponding to first bits ofthe received data on the inphase component a)ds and the quadrature is component axis is Provided as -x for the received C111a X. if the modulation type of a tr-ansaxitter terminal i3 64-QA.M.

5. A method of claim 1, wherein in step (a), the region division value of data cOrresPOUding TO sewnd bit$ of the received dam on the inphase component &,es jMd the quadrature component a.,ds is provided as - I x'+4 for the receivcd data x, ifthe modulation type Of a Mnsmittor t=zinal is 64-QANI.

6- A method of claim 1, wherein in ftp (a), the region division value of dam corresponding to third bits ofthe received data On the inPhase ComponeW axis aiad the quadrature cOmPOn=t Wds is provided as - I x 1 -4 for the received data x, if x is small 141 if er than and the modulation type of a twumitter, -terminal is 64.QAM.

7. A method of claim 1, wherein in step (&), the region division value of a data, c0"I-SPOndingto third bits ofthe received data On the inPhase component axis and tize q=katme component axis is provided as -! x 1' +6 for the received data x, if x is greater than!'41 and if the modulation type of a transmitter ter=dn4 is 64-QAM.

S. A metbod of claim 1, wherein in step (a), the region division value of a data Corresponding to respective axes Of the received data is provided as x for the received data x if the modula4on tvpe of a transmitt= terminal is QPSY. and if bits of -1, 0 on the inphase component axis aud the quadrarjre WMponent axis are r3apped to. 1, 1.

IL9

9. A method of claim 1, wherein in step (b), the MSE is cWculated by a 6= wds averaging to generate the CST.

10. A method of claim 1, wherein step (b) inchxIes:

squaring a difference between the received data. and a -received pilot information at the received pilot position; time axis avera1mg the squared difterence to calculate an MSE at each frequ=cy carrier; inversing the calculated MSE; averaging the inversed MSE in a frequency region to calculate an overall SNR and nOM2HAng the SNR. and linearly interpolating the normalized SNR to generate a CSI for each canier position.

11. A method of claim 1, wherein step (c) includes setting the CSI to one of eitlh= a first preset value or a value greater the CST if the CSI is smaller I= a second preset value, prior to multiplying the CST to the region division value.

12. A method of claim 1, wherein step (c) includes makiu- a soft decision of the CSI to a a first preset value if the CSI is smaller than a second preset value, prior to multiplying the CSI to the region division value.

13- A device for receiving and demapping a digital video from an equali2ing unit comprising. a region detecting unit linearly dividing regions of inpia data values on an inphase component axis and a quadratme component axis, rtspectiveIr, a channel state information generating unit calculating an MSE from the input data to generate a channel state information (CSI); and a decisicm and quantizing unit multiplying the CS I from the charmel state information genmvjng unit to a value divided by each region at the region deteedng unit, and quantizing a diffemce of the multiplied values.

14. A device of claim 13, wherein the ebamel state irdorniation gencratinger unit calculates the MSE by ti=e Ws averaging.

15. A device of claim 13, wherein jc channel state infonnation generating unit includes; an AMSE calculating unit squaTing a difference between the input data and a pilot information at a pilot position, and time aids averaging the squared diff- ercnce to calculate an NdSE at each frequzncy carrier, an inverting =ft obtsbing an inversed value of an output of the MSE calculating unit, a normalizing unit averaging outputs of tbue inverting unit in a fivquency region to calculate an ovemll. M. and dividing.an output of the inverting =A by the overall SNR to normalize the ouW of the inverting unit, and an interpolatiij urdt lincarly interpolamAg the nonralized data to predict a CSI at every canier position.

16- A device of claim 15, whereinthe MSE calculatingurt carries out a calculation US4 the following equation, ZL MSE(ek) pk - Ck 1) (Pkk ± n) JA n,t hk k' where k denotes a ftequency carrier index, N denotes an output of a (k)th carrier in the equalizing unit, pk denotes a refren= value, h. denotes an impulse reWme of a channeL Wk denotes a predicted value of fu hb, nk denotes a noise componera and 0 denotes a time axis avemging process.

17. A device of claim 15, wberein the MSE calculating unit cwTies out a calculation using the following equation, MSE(ek) - (Cims[ej?) whM k denotes a frequency carrier index, ek denotes an output of a (k)th carrier in tT= equalizing unit, img[ej)' denotes a quadraturt component of a,, and () denota a time axis averaging proms.

I S. A device of claim 13, wherein the decision and quantizing unit provides one ofeither a first preset value or a value greaterthanthe CSIifthe CSI is smalle: than a second preset value prior to multiplying the CSI to the region division value.

19. A de,icv of claim 13, wherein the decision and quandzing unit makes a soft decision of ft CSI to a first preset value if the CSI is smaller than a second preset value, prior to 22 multiplying the CSI to the region divisioa value.

20. A DVB-T reception rstern comprising:

a processin,& unit digitizing and convmtng a received data into a complex 4W having a quadrature component; a ftequency correcting unit correcting a fiequency offset in the complex signal, utilizing an auto ftaquency control signal; a FFT unit subjectirig the output of the fitquency correcting unit to FFT with respect to -a starting point provided by a timing s=hronizing unit; a pilot extmcting unit extracting a pilot signal from the output of the FFT unit; an oquaUzcr including a chawel jtdictor comparing a channel predictor to generate a cha=el impulse respoase, and a divider dividing the M si-pal by the channel impulse M-Spowe; a demapper receiving the output from the divider and linearly dividing regions of input data values on an inphase component axis and a quadrature component axis, respectively, calculating ac MSE from the output of the divider to generate a charm I state information (CSI) using the pilot sipal, and multiplying the CSI from the channel state informa&n generating unit to a vWue divided by eacb region at the region detecting unit, and quantizing a diffemce of the multiplied values; and a decoding unit decoding the quantized dara to display the input data.

23

21. A method for demapping a received digital video data substantially as herein described and with reference to figures 4 to 11 of the accompanying drawings.

22. A device for receiving and demapping a digital video from an equalizing unit substantially as herein described and with reference to figures 4 to I I of the accompanying drawings.

24