(19)
(11)EP 2 070 236 B1

(12)EUROPEAN PATENT SPECIFICATION

(45)Mention of the grant of the patent:
14.08.2019 Bulletin 2019/33

(21)Application number: 07838118.3

(22)Date of filing:  12.09.2007
(51)Int. Cl.: 
H04L 1/00  (2006.01)
H04W 52/20  (2009.01)
H04L 1/20  (2006.01)
H04W 52/44  (2009.01)
(86)International application number:
PCT/US2007/019856
(87)International publication number:
WO 2008/039303 (03.04.2008 Gazette  2008/14)

(54)

DISCONTINUOUS TRANSMISSION (DTX) DETECTION USING A TURBO DECODER GENERATED SIGNAL METRIC

DETEKTION EINER DISKONTINUIERLICHEN ÜBERTRAGUNG (DTX) UNTER VERWENDUNG EINER VON EINEM TURBO-DECODER ERZEUGTEN SIGNALMETRIK

DÉTECTION DE TRANSMISSION DISCONTINUE (DTX) À L'AIDE D'UN SIGNAL MÉTRIQUE GÉNÉRÉ PAR UN TURBO-DÉCODEUR


(84)Designated Contracting States:
AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HU IE IS IT LI LT LU LV MC MT NL PL PT RO SE SI SK TR

(30)Priority: 26.09.2006 US 526725

(43)Date of publication of application:
17.06.2009 Bulletin 2009/25

(73)Proprietor: Nokia of America Corporation
Murray Hill, NJ 07974-0636 (US)

(72)Inventors:
  • DOMINIQUE, Francis
    Rockaway, NJ 07866 (US)
  • KONG, Hongwei
    Denville, NJ 07834 (US)

(74)Representative: Nokia EPO representatives 
Nokia Technologies Oy Karaportti 3
02610 Espoo
02610 Espoo (FI)


(56)References cited: : 
WO-A-2007/021481
US-A1- 2004 153 950
US-A1- 2003 142 728
US-A1- 2004 240 529
  
      
    Note: Within nine months from the publication of the mention of the grant of the European patent, any person may give notice to the European Patent Office of opposition to the European patent granted. Notice of opposition shall be filed in a written reasoned statement. It shall not be deemed to have been filed until the opposition fee has been paid. (Art. 99(1) European Patent Convention).


    Description

    BACKGROUND OF THE INVENTION


    Field of the Invention



    [0001] Example embodiments of the present invention are generally related to detection of discontinuous transmission frames in transmitted data, and to a method of generating a signal metric for use in DTX detection.

    Description of Related Art



    [0002] Third generation wireless standard 3GPP2-CDMA2000-1x is designed for both voice and data applications. Typically, transmission from a base station to a mobile station in a wireless communication system is known as a forward link, and transmission from the mobile station to the base station is known as a reverse link. When a system is used for data applications, reverse link channels required to support the application usually involve a dedicated control channel (R-DCCH), which is used to transmit control information, and a supplemental channel (R-SCH), which is used to transmit data. These channels are in addition to a reverse link pilot channel, which is always transmitted.

    [0003] Due to the bursty nature of data applications, if the transmitter signal is switched on only during periods of data input, the duty cycle of the mobile station can be cut to less than 50 percent in some applications. Thus to extend mobile station battery life and to reduce interference to other users, discontinuous transmission (DTX) can be used. DTX is a method of momentarily powering-down, or muting, a mobile station when there is no data input to the transmitter. A mobile station on its own discretion decides whether to send a packet of data to the base station on a frame-by-frame basis. The mobile station decides not to send a packet of data to extend the battery life of the mobile station battery life and reduce interferences in a radio environment. DTX is used when there is no data to transmit on either channel. In other words, no signal is actually transmitted during DTX frames of a particular channel.

    [0004] In the conventional art, a mobile station (or user equipment (UE)) does not notify a base station that it has sent a frame without any symbols (data), i.e., a DTX frame. The base station (or Node-B) makes that determination on its own.

    [0005] An issue with DTX transmission is its impact on power control. A base station receives a checksum value, which is typically included at an end of a frame. Cyclic redundancy checking (CRC) checksum at a base station receiver is used to drive an outer-loop power control, so that a pre-defined frame error rate (FER) may be achieved. As explained above, the base station does not know that the mobile station has sent a frame without any data, so it processes the frame as if there is data transmitted. This may result in a CRC error since no signal is actually transmitted in that frame. This false CRC error may drive up the outer-loop power control target, which in turn increases interference level to other users and wastes power on a mobile station transmitter. Therefore, a base station receiver must detect whether a DTX frame is present, so that an outer-loop power control can either ignore a data frame CRC report or uses some other metric, (such as pilot frame error defection) to drive the outer-loop power control.

    [0006] Another type of checksum error may occur when a transmitted frame becomes distorted during transmission due to poor channel conditions. Here, the base station transmits a frame but the transmitted frame is not properly received by the base station. This type of error is known as an "erasure."

    [0007] FIGS.1A and 1B is a block diagram to illustrate Reverse link Dedicated Control Channel (R-DCCH) or Reverse link Supplement Channel (R-SCH) processing employing a conventional DTX detector. The blocks shown for the transmitter 100 at the UE and blocks at base station receiver 150 represent processing functions performed by software routines which are iterated by respective processors at the UE or Node-B respectively.

    [0008] Referring to FIG. 1A, at the UE transmitter 100, a data packet or frame (i.e., DCCH, and/or SCH data) is appended with CRC bits at CRC append unit 105, forward error code (FEC) encoded at FEC coder 110, rate adjusted at rate matching unit 115, interleaved at interleaver 120 and weighted by gains at gain unit 135 to achieve certain power levels. The pilot channel is also weighted by gains at gain unit 140 to achieve certain power levels and then spread by an orthogonal Walsh code at orthogonal spreading unit 140. The two channels are then combined (code-division multiplexed) at multiplexer 145. The multiplexed signal may be scrambled and filtered by a shaping filter (not shown) before being modulated to RF (not shown for purposes of clarity) and sent through the propagation channel 147 to the base station (Node-B) receiver 150.

    [0009] At the Node-B receiver 150, the received signal 148 first passes a matched filter (not shown for clarity) and is sent to an R-DCCH/R-SCH despreader/demodulator to generate soft symbols for further processing by blocks such as decoder 176 to recover the transmitted data from the frame. The received signal 148 is additionally received by a pilot channel processor 155, which separates the pilot channel from other channels based on its Walsh code and generates channel estimates (shown at 157) and noise energy (shown at 158). The channel estimates 157 are transmitted to the R-DCCH or R-SCH despreader & demodulator 160 to generate the soft symbols (shown at 165) for further processing in an R-DCCH/ R-SCH post processor 170 and a DTX detector 180. The noise energy 185 is used for DTX detection on the corresponding data frame by the DTX detector 180.

    [0010] The R-DCCH or R-SCH post-processing by the R-DCCH or R-SCH post processor 170 may be the reverse processing of that performed at the UE transmitter side 100. The soft symbols 165 output from the R-DCCH /R-SCH despreader & demodulator 160 are de-interleaved at de-interleaver 172, rate de-matched at rate de-matching unit 174, decoded at decoder 176, and CRC checked at CRC check unit 178 to output the frame data and/or determine a CRC pass/fail.

    [0011] The DTX detector 180 calculates a signal energy in the received frame by accumulating L2-norms in accumulator 184. The L2-norms are determined by a L2-norm calculation unit 182 based on the generated soft symbols 165. Assuming for example that the complex output signal is z=a+jb, its L2-norm is given by L2(z)=a2+b2. The L2-norms are this accumulated over the frame interval in accumulator 184 to output the signal energy.

    [0012] The detector 180 then calculates the signal-to-noise energy ratio (SNR) based on the noise energy 158 received from the pilot channel processor 155 and the determined signal energy from accumulator 184 at SNR calculation unit 186. The SNR value is then sent to a comparator 186. If the comparator 188 determines that the SNR is less than some pre-defined threshold 188, the base station receiver 150 determines that the frame is a DTX frame, (DTX On), or not (DTX Off), respectively.

    [0013] In the conventional DTX detector 180 of FIG. 1B, DTX detection performance is not satisfactory for short data frames (5ms R-DCCH, or R-SCH with low data rates). In the conventional detector of FIG. 1B, energy is estimated prior to decoding. Therefore, to remove modulation, soft symbols must be squared or the absolute value of the soft symbols must be determined (at L2-norm calculation unit 182) prior to accumulation at accumulator 184 to generate the signal energy. The conventional DTX detector 180 also cannot accurately distinguish whether a checksum error was caused by an erasure or a DTX frame. For larger data frames, e.g., R-SCH with very high data rates, especially if the detector 180 is to be implemented in Digital Signal Processing (DSP) or Field-Programmable Gate Array (FPGA), accumulation typically takes too long.

    [0014] In the DTX detector, according to the patent application US 2004/153950 A1 (TAPANINEN JUKKA ET AL) 5 August 2004 (2004-08-05), the cumulative metric of a Viterbi decoder is used for determining DTX detection.

    [0015] In another DXT detector, according to the patent application US 2003/142728 A1 (LIN YU-CHUAN) 31 July 2003 (2003-07-31), a metric that is based on the received energy per bit to noise power ratio is used for the DTX determination on a frame by frame basis.

    SUMMARY OF THE INVENTION



    [0016] The invention is set out in the appended claims.

    [0017] The embodiment of the present invention is directed to a method of generating a signal metric for use in detecting whether a transmitted data frame from a mobile station is a discontinuous transmission (DTX) frame. In the method, soft symbols are generated from a received signal carrying the transmitted data frame at a base station receiver, and the soft symbols are decoded in a turbo decoder of the receiver to generate the signal metric to be used for DTX detection.

    BRIEF DESCRIPTION OF THE DRAWINGS



    [0018] The present invention will become more fully understood from the detailed description given herein below and the accompanying drawings, wherein like elements are represented by like reference numerals, which are given by way of illustration only and thus are not limiting of the example embodiments of the present invention.

    FIGS. 1A and 1B are block diagrams of R-DCCH or R-SCH processing with a conventional DTX detector.

    FIG. 2 is block diagram illustrating DTX detection for a convolutionally-coded data frame according to an example.

    FIG. 3 is block diagram illustrating DTX detection for a turbo coded data frame according to the embodiment of the invention.


    DETAILED DESCRIPTION OF THE EXAMPLE EMBODIMENTS



    [0019] FIG. 2 is block diagram illustrating DTX detection for a convolutionally-coded data frame according to an example. As in FIG. 1A, processing on the UE transmitter side is the same, and element numbers for the corresponding processing functions in FIG. 2 are the same as FIG. 1B unless otherwise indicated.

    [0020] Referring to FIG. 2, the received signals 148 are processed initially as described in FIG. 1, with channel estimates 257 being input to the R-DCCH/R-SCH despreader & demodulator 160 to output soft symbols, and with the noise energy being extracted at pilot channel processor 155 to be sent to SNR calculation unit 186. However, instead of the soft symbols from the R-DCCH/R-SCH despreader & demodulator 160 being sent to both the DTX detector 280 (L2-norm calculation unit 182) and the R-DCCH/R-SCH post processor 270, soft symbols 265 are only input to the post processor 270. The soft symbols are de-interleaved (at 172) and rate de-matched (at 174) before being decoded by a Viterbi decoder 276. A Viterbi decoder 276 is used because the FEC coder used to encode the data frame at the transmitter 100 was convolutional coder, thus generating a convolutionally-encoded data frame for transmission.

    [0021] Further, the arrangement shown in the block diagram of FIG. 2 precludes the need for an accumulator 184 in the DTX detector 280; an accumulator 184 is not required to generate the signal energy used for the SNR calculation at 186.

    [0022] The Viterbi-decoder 276 decodes the soft symbols to recover the transmitted data from the frame. However, unlike FIG. 1B, the Viterbi decoder 276 generates a signal energy metric 285 ("signal metric") obtained at the last Viterbi decoding stage, which is sent to the DTX detector 280. In the DTX detector 280, an L2 norm is calculated for this signal energy metric in the L2-norm calculation unit 182.

    [0023] This signal energy metric 285 is referred to as a "final winning path metric" of the Viterbi decoder 276. In 3GPP and 3GPP2, this final winning path metric represents a path metric which has a final state of 0 at the last stage of the decoding process in the Viterbi decoder 276. This final winning path metric 285 is used by the L2-norm calculation unit 182 to determine a signal energy value for the received data frame.

    [0024] As discussed above, a soft symbol may be a complex signal, i.e. z=a+jb, thus the L2 norm may be represented as L2(z)=a2+b2, as is known, the squared amplitude of the final winning path metric 285 . Noise energy 285 from the pilot channel processor 155 and the signal energy from 182 are input at SNR calculation unit 286 to calculate a signal to noise ratio (SNR). Comparator 188 compares the SNR value with a given threshold (DTX threshold value). If the SNR is less than the threshold, the received frame is determined to be a DTX frame.

    [0025] Therefore, the signal input into the DTX detector 280 is a final winning path metric at the last stage of the Viterbi decoder 276. In both 3GPP and 3GPP2, the final state of the winning path is 0 because the convolutional code defined in 3GPP and 3GPP2 starts and ends with an all-zero state. This is due to adding tail bits (zero) to a data block. At the last stage of the Viterbi decoder 276, the path with state 0 as its final state has a metric which represents a coherently combined signal amplitude over an entire code block. No additional processing is required to derive this metric, as the metric is available after the Viterbi decoding process. In other words, a final winning path metric has to be calculated, regardless of whether a frame is DTXed or not. Similarly, a final winning path metric has to be calculated, regardless of whether DTX detection in the Node-B receiver 150 is present or not.

    [0026] The signal metric 285 may yield improved DTX detection performance as compared to using soft symbols, because the uncertainty of data bits in a data frame is removed after Viterbi decoding. Also using the signal metric 285 for DTX detection may be desirable because the DTX detector 280 only has to calculate an L2-norm once every data frame, while the conventional DTX detector 180 has to calculate L2 norms 384 times for a 5ms R-DCCH frame and 1536 times for 20ms R-DCCH frame, depending on a data rate of a R-SCH frame. Additional complexity savings may be achieved as there is no need for accumulation operations to determine the signal energy for the SNR calculation.

    [0027] FIG. 3 is block diagram illustrating DTX detection for a turbo coded data frame according to the embodiment of the invention. As in FIG. 1A, processing on the UE transmitter side is the same, and element numbers for the corresponding processing functions in FIG. 2 are the same as FIG. 1B unless otherwise indicated. Unlike FIGS. 1B and 2, the DTX detector 320 in FIG. 3 includes an accumulator 380 prior to the L2-Norm calculation unit 182, and the post processor 370 includes a turbo decoder 376 instead of the Viterbi decoder 276. A turbo decoder 376 is used because the FEC coder used to encode the data frame at the transmitter 100 was a turbo encoder, thus generating a turbo-encoded data frame for transmission. Like the Viterbi decoder 276, the turbo decoder 376 also generates a signal metric for input to the DTX detector 380, albeit a different signal metric.

    [0028] The signal metric input into the DTX detector 380 from the post processor 370 is a final turbo decoded Log-Likelihood Ratio (LLR) for all systematic bits in a data frame being processed by the turbo decoder 376. As is well known in the art, turbo code is systematic code, where the coded sequence consists of systematic bits and parity bits. A turbo decoder computes LLRs for all systematic bits from the received soft symbols corresponding to all coded bits, including systematic bits and parity bits.

    [0029] The DTX detector 380 accumulates LLR amplitudes of all systematic bits over the entire data frame in accumulator 382 to output a sum. The L2 norm calculation unit 182 squares the sum to output a signal energy value for the transmitted data frame that is to be used in the SNR calculation. The functions of the comparator 188 are the same as described above and hence are omitted for brevity.

    [0030] The sum value determined at 182 represents a signal energy which may yield improved DTX detection performance as compared with using soft symbols. From a complexity point of view, the L2-norm calculation is required only once every data frame, while the conventional DTX detector has to do the L2-norm calculation between 1536 times and 12288 times per frame, depending on the data rate of the R-SCH frame. The number of accumulation operations by accumulator 382 is reduced by at least 75% (excluding repetition due to rate-matching) as compared to accumulator 182 of the conventional DTX detector 180, and with a code rate of ¼ (since accumulation is on systematic bits only).

    [0031] Therefore, using certain metrics from a R-DCCH or a R-SCH post-processing decoder 276/376 instead of soft symbols may improve DTX detection performance for data frames of any size (short and long) with reduced complexity.


    Claims

    1. A method of detecting at a base station whether a transmitted data frame from a mobile station, UE, is a discontinuous transmission, DTX, frame, comprising:

    generating, from a signal (148) carrying the frame that is received by the base station receiver (150), a signal metric (380) corresponding to the transmitted data frame in a decoding operation used to decode the data frame,

    extracting, in a pilot channel processor (155), a noise energy value from the transmitted data frame,

    determining a signal energy of the transmitted data frame based on the signal metric, and

    determining whether the transmitted data frame is a DTX frame using a calculated signal-to-noise ratio based on the determined signal energy and on the extracted noise energy value,

    wherein the method of detecting is characterised by

    generating the signal metric (380) includes:

    generating soft symbols at the base station from a received signal (148) carrying the transmitted data frame, and

    decoding the soft symbols in a turbo decoder (376) to generate the signal metric to be used for DTX detection, wherein decoding includes:

    computing log-likelihood ratio, LLR, amplitudes for all systematic bits in the transmitted data frame, and

    determining a final turbo decoded Log-Likelihood Ratio, LLR, for each of the systematic bits in the data frame as the signal metric for DTX detection;

    inputting the signal metric into an accumulator (382) of LLR amplitudes of all systematic bits only, over the entire data frame,

    wherein the signal energy of the transmitted data frame is determined as an energy value of the output of the accumulator.


     
    2. The method of claim 1, wherein determining whether the transmitted data frame is a DTX frame includes:
    comparing the calculated signal-to-noise ratio to a given threshold value to determine if the transmitted data frame is a DTX frame.
     


    Ansprüche

    1. Verfahren zum Detektieren an einer Basisstation, ob ein von einer Mobilstation, UE, übertragener Datenrahmen ein diskontinuierlicher Übertragungs-DTX-Rahmen ist, umfassend:

    Erzeugen einer Signalmetrik (380) aus einem Signal (148), das den Rahmen trägt, der von dem Basisstationsempfänger (150) empfangen wird und dem übertragenen Datenrahmen in einem Dekodiervorgang entspricht, der zum Dekodieren des Datenrahmens verwendet wird,

    Extrahieren eines Rauschenergiewerts aus dem übertragenen Datenrahmen in einem Pilotkanalprozessor (155),

    Bestimmen einer Signalenergie des übertragenen Datenrahmens basierend auf der Signalmetrik, und

    Bestimmen, ob der übertragene Datenrahmen ein DTX-Rahmen ist, unter Verwendung eines berechneten Signal-Rausch-Verhältnisses basierend auf der bestimmten Signalenergie und dem extrahierten Rauschenergiewert, wobei das Verfahren zum Erfassen gekennzeichnet ist durch:
    Erzeugen der Signalmetrik (380) beinhaltet:

    Erzeugen von weichen Symbolen an der Basisstation aus einem empfangenen Signal (148), das den übertragenen Datenrahmen trägt, und

    Dekodieren der weichen Symbole in einem Turbodekodierer (376), um die Signalmetrik zu erzeugen, die für die DTX-Detektion verwendet werden soll, wobei das Dekodieren beinhaltet:

    Berechnen des "Log-Likelihood" Quotienten, LLR, von Amplituden für alle systematischen Bits in dem übertragenen Datenrahmen, und

    Bestimmen eines endgültigen turbodekodierten "Log-Likelihood" Quotienten, LLR, für jedes der systematischen Bits im Datenrahmen als Signalmetrik für die DTX-Detektion;

    Eingeben der Signalmetrik in einen Akkumulator (382) von LLR-Amplituden aller nur systematischen Bits über den gesamten Datenrahmen,

    wobei die Signalenergie des übertragenen Datenrahmens als ein Energiewert des Ausgangs des Akkumulators bestimmt wird.


     
    2. Verfahren nach Anspruch 1, wobei das Bestimmen, ob der übertragene Datenrahmen ein DTX-Rahmen ist, beinhaltet:
    Vergleichen des berechneten Signal-Rausch-Verhältnisses mit einem gegebenen Schwellenwert, um festzustellen, ob der übertragene Datenrahmen ein DTX-Rahmen ist.
     


    Revendications

    1. Procédé de détection au niveau d'une station de base pour savoir si une trame de données transmise à partir d'une station mobile, UE, est une trame de transmission discontinue, DTX, comprenant :

    la génération, à partir d'un signal (148) acheminant la trame qui est reçue par le récepteur de station de base (150), d'une métrique de signal (380) correspondant à la trame de données transmise dans une opération de décodage utilisée pour décoder la trame de données,

    l'extraction, dans un processeur de canal pilote (155), d'une valeur d'énergie de bruit à partir de la trame de données transmise,

    la détermination d'une énergie de signal de la trame de données transmise sur la base de la métrique de signal, et

    la détermination pour savoir si la trame de données transmise est une trame DTX utilisant un rapport signal sur bruit calculé basé sur l'énergie de signal déterminée et sur la valeur d'énergie de bruit extraite,

    dans lequel le procédé de détection est caractérisé par la génération de la métrique de signal (380) qui comprend :

    la génération de symboles logiciels au niveau de la station de base à partir d'un signal reçu (148) acheminant la trame de données transmise, et

    le décodage des symboles logiciels dans un décodeur turbo (376) pour générer la métrique de signal à utiliser pour une détection DTX, dans lequel le décodage comprend :

    le calcul d'amplitudes d'un rapport de vraisemblance logarithmique, LLR, pour tous les bits systématiques dans la trame de données transmise, et

    la détermination d'un rapport de vraisemblance logarithmique, LLR, final turbo décodé pour chacun des bits systématiques dans la trame de données en tant que métrique de signal pour une détection DTX ;

    l'entrée de la métrique de signal dans un accumulateur (382) d'amplitudes LLR de tous les bits systématiques uniquement, sur toute la trame de données, dans lequel l'énergie de signal de la trame de données transmise est déterminée en tant que valeur énergétique de la sortie de l'accumulateur.


     
    2. Procédé selon la revendication 1, dans lequel la détermination pour savoir si la trame de données transmise est une trame DTX comprend :
    la comparaison du rapport signal sur bruit calculé à une valeur seuil donnée pour déterminer si la trame de données transmise est une trame DTX.
     




    Drawing















    REFERENCES CITED IN THE DESCRIPTION



    This list of references cited by the applicant is for the reader's convenience only. It does not form part of the European patent document. Even though great care has been taken in compiling the references, errors or omissions cannot be excluded and the EPO disclaims all liability in this regard.

    Patent documents cited in the description