JP2005198223A - Multi-user detection receiver for packet transmission in multi-carrier - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a receiver which attains the elimination of an interference between users caused by reception timing deviation exceeding GI and the improvement of channel estimation accuracy in a packet switched multi-carrier transmission system in an upstream line of mobile communication. <P>SOLUTION: The multi-user detection receiver includes: a timing reproducer 2 for reproducing timing from a reception signal; a channel estimator 3 for estimating a channel impulse response; a multi-user detector 4 for detecting the reception signal using the timing and the channel impulse response; a decoder 6 for performing error-correcting decoding upon a detection signal; a packet error detector 8 for detecting an error of a packet; a repetition controller 10 which performs control to repeat a reception process if a packet error is detected; and a replica generator 12 which generates a replica of a transmission signal using the result of decoding. In repetition control, the replica of the transmission signal is used for allowing the multi-user detector 4 to perform interference cancel and the channel estimator 3 to decision-oriented channel estimation. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  The present invention relates to a multiuser detection receiver for packet transmission in a multicarrier.

  In the next generation mobile communication system that realizes high-speed signal transmission, the spread code is applied to the OFDM system that implements the multi-carrier transmission system using the fast Fourier transform (FFT) and the inverse fast Fourier transform (IFFT), and the OFDM subcarrier. OFCDM (or called MC-CDMA) scheme that uses and spreads the signal has been studied.

  Since these methods use FFT and IFFT, it is necessary to ensure continuity of signals within the Fourier transform interval. In a multipath environment, a delay time occurs in the received signal. Therefore, the continuity of the signal is ensured in the receiver by inserting the rear part of the symbol as a guard interval (GI) before the symbol to generate a transmission signal. However, the GI needs to be set longer than the delay time. For this reason, adoption is considered in the downlink where the signals of each user are received simultaneously.

  On the other hand, since the distance to the base station differs depending on the position of each user on the uplink, a reception timing deviation may occur and the deviation may be greater than or equal to GI. In such an environment, since the continuity of the signal is not ensured, inter-carrier interference (ICI) that occurs due to the loss of orthogonality between subcarriers in the symbol at the current time and the symbol at the previous or next time Leaky intersymbol interference (ISI) occurs. In the uplink, these interferences occur between users.

Therefore, MC-DS-CDMA, which is a transmission scheme that is resistant to interference generated between users in the uplink, has been studied as a promising candidate (see Non-Patent Document 1, for example). However, studies are also being conducted on the application of OFDM and OFCDM to the uplink for the purpose of reducing component costs and miniaturizing terminals by using the same transmission method for the uplink and downlink.
S. Suwa et al., “Performance comparison between MC / DS-CDMA and MC-CDMA for reverse link broadband packet wireless access” Proc. IEEE VTC 2002fall, pp. 2076-2080, 2003

As a method for compensating the reception timing deviation between users, there is known a method of preparing a GI having a length that can allow the remaining timing deviation in a situation where a certain level of reception timing synchronization is established (for example, Patent Document 2).
S. Tsumura et al., “Design and performance of quasi-synchronous multi-carrier CDMA system” Proc. IEEE VTC 2001 fall, pp. 843-847, 2001

  There is also a method for realizing high-accuracy reception timing control by performing packet transmission / reception between the base station and the terminal several times and feeding back the reception timing deviation measured by the base station to the terminal side.

However, the conventional method has the following drawbacks.
(1) In the method of preparing a GI having a length that allows the timing deviation, the transmission efficiency decreases as the GI length increases.
(2) The method of preparing a GI having a length that allows the timing deviation is greatly deteriorated due to interference unless reception timing synchronization is established to the extent that the GI length allows the timing deviation.
(3) The method of establishing a high-accuracy reception timing by feedback control requires a large number of packets for control, so that transmission efficiency decreases.
Considering the above points, it is desirable that interference generated by a reception timing deviation exceeding the GI can be removed only by the receiver without reducing the transmission efficiency of the system.

  In addition, since the signal transmitted from the terminal is received on the uplink, it is difficult to obtain the received signal power, so that there is a problem that the estimation accuracy of the channel impulse response of each user is significantly deteriorated.

  The present invention has been made in view of such problems, and in an uplink multi-carrier transmission system in a mobile communication system, interference generated by disruption of orthogonality between users due to reception timing deviation exceeding GI is achieved. An object of the present invention is to provide a receiver that can be eliminated and that can improve channel estimation accuracy.

  The multi-user detection receiver for packet transmission in the multi-carrier of the present invention has L (L is a positive integer) antennas for receiving packets transmitted by U (U is a positive integer) users by the multi-carrier transmission method. A timing regenerator that estimates the reception timing of the user, a channel estimator that estimates the channel impulse response of the user, a multi-user detector that detects a received signal and outputs a detected signal, and A decoder that performs error correction decoding after deinterleaving the detection signal, a bit determination unit that determines a received bit from the decoding result, a packet error detector that detects an error in the received bit of the packet, and a transmission from the decoding result A replica generator that generates a replica of the signal and a control that repeatedly operates all or part of these components. Those having a repetitive controller that performs.

  The timing regenerator of the present invention estimates the reception timing deviation of the user from the cross-correlation value between the pilot signal and the received signal.

  In addition, the channel estimator of the present invention generates a replica of the received signal using the transversal filter as an input, and generates a replica of the received signal, and uses the least square method to minimize a difference from the received signal. By estimating the tap weight coefficient of the Versal filter, the channel impulse response of the user is estimated, or the channel estimation is similarly performed with the replica of the transmission signal as an input.

  The timing regenerator of the present invention determines the necessary number of taps of the transversal filter in the channel estimator from the cross-correlation value.

  The multiuser detector of the present invention derives the transfer function of the user using the timing deviation and the channel impulse response, and weights and synthesizes the received signal after Fourier transform using the transfer function. In the turbo interference canceller using the replica of the transmission signal, a replica of the received signal other than the desired user is generated and subtracted from the received signal. After conversion, weighted synthesis is performed to output a detection signal.

  Further, the decoder of the present invention performs error correction decoding by maximum a posteriori probability decoding after deinterleaving the detection signal, and outputs a log likelihood ratio between the information bit and the encoded bit.

  Further, the decoder of the present invention combines the detection signal of the previous packet received in error and the detection signal retransmitted this time, and decodes the maximum posterior probability when the packet is retransmitted. It is.

  The bit determination unit of the present invention determines a received bit from the sign of the log likelihood ratio of the information bit.

  The packet error detector of the present invention detects an error in the packet by decoding a cyclic code for the received bit determined using the cyclic code added to the packet.

  Further, the replica generator of the present invention calculates an expected value of the user's modulation signal using the logarithmic likelihood ratio of the encoded bits, and uses the expected value of the modulation signal as a modulation signal. A replica of the transmission signal is generated by performing the same processing.

  Further, the repeat controller of the present invention repeats either all or a part of the receiver configuration by a predetermined number of times or repeats until no error is detected in the packet of all users. In both cases, when there is no packet error for each user, the reception process for that user is not repeated.

  According to the present invention, it is possible to eliminate interference caused by a reception timing deviation exceeding GI, and to improve channel estimation accuracy.

The present invention has the following effects.
According to the multi-user detection receiver of the first aspect of the present invention, in uplink multi-carrier transmission in packet switching, interference generated when there is a reception timing deviation exceeding the GI between users and accuracy of channel estimation can be improved. Improvements can be made.

  Hereinafter, the best mode for carrying out the present invention will be described.

  First, the best mode for carrying out the first invention according to the multiuser detection receiver will be described. FIG. 1 shows a basic configuration of a multiuser detection receiver. A signal transmitted by U users is received by L antennas. The received signal is input from the antenna terminal 1. The timing regenerator 2 estimates the reception timing of each user from the received signal. The output of the timing regenerator 2 is connected to the channel estimator 3 and the multiuser detector 4, and the estimated value is input to each. The channel estimator 3 estimates the channel impulse response of each user from the received signal and the reception timing. The estimated value is input to the connected multi-user detector 4. The multiuser detector 4 detects the received signal using the reception timing and channel impulse response of each user, and outputs the detected signal of each user. The detected signal of each user is deinterleaved by the deinterleaver 5 and is decoded by the decoder 6 with maximum a posteriori probability (MAP). The bit determiner 7 determines the received bit from the output decoding result. Further, the packet error detector 8 detects the error of the received bit in the packet by decoding the cyclic code added to the packet. When a packet error is detected, the repeat controller 10 shifts to repeat processing. In the iterative process, after the interleaver 11 interleaves the decoding result, the replica generator 12 generates and outputs a replica of each user's transmission signal. The output transmission signal replica is input to the channel estimator 3 and the multiuser detector 4. During the iterative processing, the channel estimator 3 estimates the channel impulse response again using the replica of the transmission signal. The estimation result is input again to the multiuser detector 4. The multiuser detector 4 again outputs the detection signal of each user from the received signal. At that time, the turbo interference canceller 16 shown in FIG. 2 operates in the multi-user detector 4 at the time of repetition, and cancels received signals other than the desired user. Details will be described later. The detection signal of each user is similarly processed by the decoder 6. The repetitive controller 10 performs control to repeatedly operate all or a part of these configurations. The control signal of the repeat controller 10 is connected to each unit.

  From the above, according to the best mode for carrying out the present invention, the reception timing is estimated and generated using the channel impulse response that estimates the received signal replica other than the desired user based on the deviation. And since it removes from a received signal, the interference between users which generate | occur | produces by reception timing deviation can be removed. Further, estimation accuracy can be improved by performing channel estimation in the data signal section using the decoding result.

  Next, the best mode for carrying out the second invention according to the multiuser detection receiver will be described. The present invention relates to the operation of a timing regenerator. FIG. 3 shows the configuration of the timing regenerator. In the timing regenerator, a cross-correlation value between a pilot signal and a received signal that are different for each user time-multiplexed with the transmission signal of each user is calculated by the correlator 22 for the received signal of each antenna. The cross-correlation values of the antennas are calculated by the absolute value squarer 23 and the absolute value square value thereof is calculated and synthesized. The combined cross-correlation value determines the time when the maximum value is obtained by the threshold detector 24 as the reception timing of the user. Similar processing is performed for all users using the timing regenerator 21 for each user. For the user whose cross-correlation value does not exceed the predetermined threshold, it is determined that transmission is not performed, and all subsequent reception processing is not performed.

  From the above, according to the best mode for carrying out the present invention, the cross-correlation value is calculated using a pilot signal different for each user, so that the reception timing deviation of each user can be estimated. Moreover, the user who transmitted can be determined. In addition, since the cross-correlation values calculated for each antenna are combined, more accurate estimation can be performed.

  Next, the best mode for carrying out the third and fourth aspects of the multiuser detection receiver will be described. The present invention relates to the operation of a channel estimator. FIG. 4 shows the configuration of the channel estimator. Since channel estimation is performed for each antenna, it is necessary to prepare each antenna channel estimator 26 for the number of antennas. Each antenna channel estimator 26 collectively estimates the channel impulse response of each user by the least square method using a transversal filter.

  Specifically, a replica is generated for the received signal in the time domain at the time discretized based on the reception timing of each user estimated by the timing regenerator, and the absolute value square of the difference between the received signal and the replica is generated. It is estimated by the least square method by the weighting coefficient controller 27 so that the value is minimized. The input of the transversal filter is a pilot signal of each user, and when the channel impulse response is used to estimate the tap weighting coefficient, the output is a replica of the received signal. At this time, the pilot signal of each user is controlled by the input timing controller 29 and input based on the reception timing of each user estimated by the timing regenerator. Note that the number of taps of the transversal filter is determined by the threshold detector 24 based on the time width exceeding α (1> α ≧ 0) times the maximum value of the cross-correlation value for each user. The tap controller 28 changes the tap number of the transversal filter in accordance with the determined tap number.

  Further, at the time of iterative control, decision-directed channel estimation is performed using a replica of the transmission signal generated by the replica generator 12 as an input to the transversal filter. The initial value of the tap weighting coefficient is a final estimated value estimated using a pilot signal. The estimated channel impulse response is again input to the multi-user detector and used in the turbo interference canceller.

  From the above, according to the best mode for carrying out the present invention, in order to directly estimate the channel impulse response by changing the number of parameters to be estimated according to the propagation environment of each user, the transfer function in the frequency domain is The accuracy can be improved more than the estimation. In addition, since the number of parameters to be estimated can be reduced, estimation is possible even if the length of the pilot signal is short. During iterative control, channel estimation can be performed in the data signal section in addition to the pilot signal, so that the estimation accuracy can be improved. Furthermore, since the probability of the replica of the transmission signal is improved by increasing the number of repetitions, the estimation accuracy can be further improved.

  Next, the best mode for carrying out the fifth aspect of the multiuser detection receiver will be described. The present invention relates to the configuration and operation of a multi-user detector. FIG. 2 shows the configuration of the multiuser detector. The multi-user detector 4 includes a multi-user detector 14 for each user, and converts a received signal into a frequency domain by an FFT unit at an individual timing for each user according to the estimated timing deviation of each user. . Only the repetitive control operates the turbo interference canceller 16 and the received signal replica generation 15. Details will be described later.

The weighting coefficient is determined by the least square method so that the error between the combined signal obtained by performing weighted combining on the received signal after the Fourier transform and the transmission signal is minimized. Derivation of the weighting coefficient is shown in Non-Patent Document 3, for example. In OFDM, the weighting coefficient is only a transfer function, and in OFCDM, it is generated using a spreading code. The transfer function can be derived by Fourier-transforming the estimated channel impulse response of each user in consideration of individual timing for each user. The weighting coefficient is generated by the weighting coefficient generator 18, and the weighted synthesizer 19 synthesizes the received signal using the weighting coefficient.
Kenji Akita, Satoshi Suyama, Kazuhiko Fukawa, Hiroshi Suzuki, "Interference Canceller in Downlink Using MC-CDMA Transmission Power Control" IEICE Technical Report RCS2002-35, April 2002

The weighted synthesized signal is converted into a log likelihood ratio for each bit in the soft decision output detector 20. For example, Non-Patent Document 4 shows a conversion method from a combined signal to a log likelihood ratio. After the conversion, the multiuser detector 4 outputs a log likelihood ratio of bits for all users as a detection signal.
Jun Suyama, Hiroshi Suzuki, Kazuhiko Fukawa, "OFDM turbo equalization for multipath delay exceeding guard interval" IEICE Technical Report RCS2002-202, January 2003

  Further, at the time of repetitive control, the multi-user detector 14 for each user extracts only the received signal component of the desired user from the received signal by the turbo interference canceller 16 and performs weighted synthesis on the extracted received signal. Output the composite signal. The combined signal is converted into a log-likelihood ratio of bits as described above.

  The reception signal replica generator 15 generates a replica of the reception signal other than the desired user by using the replica signal of each user's transmission signal and the estimated timing and channel impulse response. Next, only the received signal component of the desired user is extracted by subtracting the received signal replica from the received signal. The received signal component of the desired user is subjected to FFT in accordance with the timing of the desired user, and weighted and synthesized by the least square method. For example, in OFDM, each subcarrier is weighted so as to be divided by a transfer function, and synchronous detection is performed. In addition, in OFCDM, despreading is simultaneously performed by weighting synthesis. In addition, when weighted synthesis is performed, the characteristics are improved by taking into account the residual error after subtracting the replica of the received signal in the received signal. For example, Non-Patent Document 4 shows a method for deriving a weighting coefficient in consideration of a residual error of replica subtraction.

  From the above, according to the best mode for carrying out the present invention, since FFT is performed at different timing for each user, signal power of a desired user can be secured and interference power can be reduced. In addition, since the received signals are weighted and synthesized with a strict transfer function of each user, it is possible to suppress interference caused by a reception timing deviation exceeding the GI between users. Further, during repetitive control, the received signal components other than the desired user are subtracted from the received signal and weighted synthesis is performed, so that interference between users can be removed. Furthermore, since it is not necessary to suppress interference from other users by performing weighted combining after removing the interference, it is possible to set a weighting factor that suppresses noise more, and to improve the S / N ratio after combining.

  Next, the best mode for carrying out the sixth, eighth and ninth aspects of the multiuser detection receiver will be described. The present invention relates to the operation of a decoder. The decoder performs error correction decoding by MAP decoding after deinterleaving the log likelihood ratio of each user's bit output from the multiuser detector. These processes are performed for each user.

  Specifically, the decoder performs MAP decoding using the log likelihood ratio of the bits output from the multiuser detector as prior information. MAP decoding is implemented by Viterbi decoding in the forward and backward logarithmic regions, and calculates each log likelihood ratio from the metric difference between the encoded bits and the information bits, and subtracts the prior information for the encoded bits. Output.

  In addition, the bit determination unit 7 determines a reception bit from the positive / negative of the log likelihood ratio of the information bit. Further, the packet error detector 8 detects an error in the packet by decoding the cyclic code for the received bit determined using the cyclic code added to the packet. When an error is detected in the packet, the log likelihood ratio of the encoded bits after decoding is output and used by the replica generator.

  From the above, according to the best mode for carrying out the present invention, in addition to the log likelihood ratio of information bits for determination, the log likelihood ratio of coded bits is also output from the encoder. Therefore, the log likelihood ratio of the encoded bits can be used for generating the replica signal in the iterative process when an error is detected in the packet. Thus, turbo interference cancellation and decision-directed channel estimation can be realized by using a MAP decoder having a soft decision output.

  Next, the best mode for carrying out the seventh aspect of the multiuser detection receiver will be described. The present invention relates to the operation of a decoder. In the decoder 6, when the packet is retransmitted, the detection signal of the previous packet received in error and the detection signal retransmitted this time are combined to perform maximum posterior probability decoding.

  From the above, according to the best mode for carrying out the present invention, the S / N ratio can be improved because the detection signal of the previous packet received in error and the detection signal of the retransmitted packet are combined. The determination reliability can be improved.

  Next, the best mode for carrying out the tenth aspect of the multiuser detection receiver will be described. The present invention relates to the operation of a replica generator. The replica generator calculates the expected value of each user's modulated signal from the log-likelihood ratio of each user's encoded bits output by the decoder, and uses the expected value of the modulated signal as the modulated signal to perform the same processing as the transmitter To generate a replica of the transmission signal.

The transmitter converts a plurality of encoded bits into a modulated signal. The conversion method differs depending on the modulation method. For example, 2 bits are converted into a modulated signal in QPSK, and 4 bits are converted into a modulated signal in 16QAM. The probability that each bit is 0 or 1 is calculated from the log likelihood ratio of the encoded bits, and the probability that each modulated signal is generated is calculated based on the probability. Furthermore, the expected value of the modulation signal can be calculated from each modulation signal and its generation probability (see, for example, Non-Patent Document 4). Next, an expected value of the modulation signal, that is, a replica of the transmission signal is generated by performing the same processing as the transmitter using the modulation signal replica as the modulation signal. For example, in OFDM, a replica of a transmission signal can be generated by IFFT of an expected value of a modulated signal and inserting a guard interval. In addition, in OFCDM, diffusion processing may be performed in addition thereto.

  From the above, according to the best mode for carrying out the present invention, the modulation signal is usually a fixed value, but in order to generate the expected value of the modulation signal from the log likelihood ratio of the encoded bits, A non-fixed, that is, soft, modulation signal can be generated. In addition, since the expected value of the modulation signal changes in an analog manner in accordance with the log likelihood ratio of the encoded bits, the characteristics can be improved when iterative processing is performed.

  Next, the best mode for carrying out the eleventh aspect of the multiuser detection receiver will be described. A control line is connected from the repeat controller to each unit in the multi-user detection receiver, and the repeat controller centrally controls the repetition of the reception process. The control is performed with a smaller number of repetitions, that is, the operation is repeated for the number of times or the operation is repeated until no error is detected in the packet of all users in the bit decision unit. In either case, when there is no packet error for each user, the receiving process for that user is not repeated.

  From the above, according to the best mode for carrying out the present invention, when there is no packet error for each user, the reception process for that user is not repeated, so a series of reception processes are efficiently repeated. be able to.

  It should be noted that the present invention is not limited to the best mode for carrying out each invention, and various other configurations can be adopted without departing from the gist of the present invention.

It is a figure which shows the basic composition of the multiuser detection receiver for packet transmission in the multicarrier by this invention. It is a figure which shows the structural example of a multiuser detector. It is a figure which shows the internal structural example of a timing regenerator. It is a figure which shows the internal structural example of a channel estimator.

Explanation of symbols

1: antenna terminal, 2: timing regenerator, 3: channel estimator, 4: multiuser detector, 5: deinterleaver, 6: decoder, 7: bit determiner, 8: packet error detector, 9: Receive bit output terminal, 10: repetitive controller, 11: interleaver, 12: replica generator, 13: pilot signal input terminal, 14: multi-user detector for Uth user, 15: received signal replica generator, 16: Turbo interference canceller, 17: FFT unit, 18: weighting coefficient generator, 19: weighting synthesizer, 20: soft decision output detector, 21: timing regenerator for Uth user, 22: correlator, 23: absolute value 2 Multiplier, 24: threshold detector, 25: threshold input terminal, 26: channel estimator for Lth antenna, 27: weighting coefficient controller, 28: tap controller, 29: input counter Timing controller

Claims (11)

Input from L (L is a positive integer) antennas that receive packets transmitted by a multi-carrier transmission method by U users (U is a positive integer) encoded and multiplexed with pilot signals; A timing regenerator for estimating the reception timing of the user; a channel estimator for estimating the channel impulse response of the user; a multi-user detector for detecting a received signal and outputting a detected signal; and deinterleaving the detected signal A decoder that performs error correction decoding later, a bit determination unit that determines a received bit from a decoding result, a packet error detector that detects an error in the received bit of the packet, and a replica of a transmission signal is generated from the decoding result Replica generator, and repetitive controller that performs control to repeatedly operate all or part of these components. It is a multi-user detection receiver. The timing regenerator according to claim 1, wherein a timing regenerator that estimates a reception timing deviation of the user from a cross-correlation value between the pilot signal and the received signal is used. User detection receiver. 2. The channel estimator according to claim 1, wherein a replica of the received signal is generated using the transversal filter as an input using a transversal filter, and the difference between the received signal and the received signal is minimized by the least square method. A channel estimator that estimates a channel impulse response of the user by estimating a tap weight coefficient of a transversal filter or similarly performs channel estimation using a replica of the transmission signal as an input is used. The multi-user detection receiver according to claim 1. The timing regenerator according to claim 1, wherein a timing regenerator that determines the required number of taps of the transversal filter in the channel estimator from the cross-correlation value is used. User detection receiver. 2. The multiuser detector according to claim 1, wherein the transfer function of the user is derived using the timing deviation and the channel impulse response, and the received signal after Fourier transform is weighted and synthesized using the transfer function. In the turbo interference canceller using the replica of the transmission signal, a replica of the received signal other than the desired user is generated and subtracted from the received signal. The multiuser detection receiver according to claim 1, wherein a multiuser detector that outputs a detection signal by weighting and combining after Fourier transform is used. The decoder according to claim 1, wherein the detection signal is subjected to error correction decoding by maximum a posteriori probability decoding after deinterleaving, and a logarithmic likelihood ratio between information bits and encoded bits is used. The multi-user detection receiver according to claim 1, wherein: 2. The decoder according to claim 1, wherein when a packet is retransmitted, the detection signal of the previous packet received in error and the detection signal retransmitted this time are combined, and maximum a posteriori probability decoding is performed. The multi-user detection receiver according to claim 1, wherein a decoder is used. The multi-user detection receiver according to claim 1, wherein the bit determination unit according to claim 1 uses a bit determination unit that determines a reception bit based on a sign likelihood ratio of the information bit. 2. The packet error detector according to claim 1, wherein an error of the packet is detected by decoding a cyclic code with respect to the received bit determined using a cyclic code added to the packet. The multi-user detection receiver according to claim 1, wherein a multi-user detection receiver is used. The replica generator according to claim 1, wherein an expected value of the modulation signal of the user is calculated using a log likelihood ratio of the encoded bits, and the expected value of the modulation signal is used as a modulation signal. The multi-user detection receiver according to claim 1, wherein a replica generator that generates a replica of a transmission signal by performing the same processing as in the first embodiment is used. 2. The repetitive controller according to claim 1, wherein either the control of repeating all or part of the receiver configuration a predetermined number of times or the control of repeating until no error of the packet of all users is detected is repeated. 2. The multi controller according to claim 1, wherein a repetitive controller is used that performs control with a small number of times and performs control so that reception processing for each user is not repeated when there is no packet error in either user. User detection receiver.

Images