JP5036655B2 - Receiving apparatus and demodulation method - Google Patents

Description

  The present invention relates to a receiving apparatus and a demodulation method for receiving and demodulating a modulated signal.

  In MIMO transmission using multi-level symbols, signal detection accuracy is greatly degraded due to the influence of inter-antenna interference and the like, and transmission characteristics are also degraded. In order to solve these degradations, it is necessary to improve the signal detection accuracy by using repeated MIMO equalization and interference cancellation. As conventional equalization / interference cancellation, for example, as described in Non-Patent Document 1 below, MIMO (Multiple Input Multiple Output) equalization and interference cancellation are repeatedly performed in order to improve signal detection accuracy. .

  In the technique described in Non-Patent Document 1 below, a QAM (Quadrature Amplitude Modulation) symbol is regarded as a hierarchy of a plurality of QPSK (Quadrature Phase Shift Keying) sub-symbols, and repeated demodulation processing is performed in parallel for each transmission antenna and each layer. Is going to. In the first processing, only MIMO equalization is performed on the received signal, and soft decision is performed in parallel for each transmission antenna and each layer. Thereafter, reliability information is obtained via an error correction decoder, and the reliability information is converted into a QPSK sub-symbol replica necessary for the next iterative demodulation.

  The replicas thus obtained are shared by the transmission antenna and layer repetition demodulation, and the repetition demodulation processing by MIMO equalization and interference cancellation is performed for each transmission antenna and layer using these replicas. In the interference cancellation processing for a desired layer for a certain transmission antenna, QAM symbols from other antennas and QPSK sub-symbols of other layers excluding the desired layer are regarded as interference components, and reception after MIMO equalization is performed. An interference component is subtracted from the signal using a replica. Similar interference cancellation processing is executed for each transmission antenna and each layer. Thereafter, the soft decision is performed again for each transmission antenna and each layer, and the above demodulation process is repeated a predetermined number of times.

“Adaptive Transmission With Single-Carrier MultilevelBICM”, Matsumoto. T, Ibi. S, Sampei. S, Thomas R, Proceedings of the IEEE Volume 95, Issue 12, Dec. 2007 pp.2354-2367

  However, according to the technique described in the above-mentioned conventional Non-Patent Document 1, replicas are used in sub-symbol units in all layers, that is, in QAM symbol units, without considering the difference in signal detection accuracy of layered QPSK sub-symbols. ing. Note that the sub-symbol hierarchical structure is characterized in that the signal determination accuracy of the QPSK sub-symbol is worse in the upper layer (assuming that the amplitude is lower in the upper layer) and the replica generation accuracy is also lower. For this reason, in the upper layer, a low-accuracy sub-symbol replica is generated and used, and as a result, there is a problem that a large error in removing interference components caused by interference cancellation and a good signal detection accuracy cannot be obtained. there were.

  The present invention has been made in view of the above, and in the case of demodulating a multilevel symbol as a hierarchical sub-symbol, it reduces the influence of an error in a layer with low signal determination accuracy, and achieves high signal detection accuracy. It is an object to obtain a receiving device and a demodulation method that can be obtained.

  In order to solve the above-described problems and achieve the object, the present invention considers one symbol of a multilevel modulated received signal as a plurality of hierarchized sub-symbols, and equalizes each hierarchized layer unit. A receiving apparatus that performs processing and interference component removal processing, instructing a layer with the highest signal accuracy as a processing target of each processing, and after completing processing based on the instructions, to the processing target layer Next, a layer with the next highest signal accuracy is added, and two layers are instructed as the next processing target in the order of the higher signal accuracy. Thereafter, until all layers in the hierarchy are processed , A control means for instructing a result of adding a layer having the next highest signal accuracy with respect to a previous processing target layer as a processing target, and a sub symbol which is a replica of a sub symbol of the processing target layer based on the instruction Sub symbol that generates replica Replica generation means, and transmission path estimation means for performing transmission path estimation based on the sub-symbol replica, performing the equalization processing based on the estimation result of the transmission path estimation means, and based on the sub-symbol replica Then, the interference component removal process is performed.

  According to the present invention, when demodulation is performed by hierarchizing multi-level modulated symbols into sub-symbols, the layer that is estimated to have the highest accuracy of replica of interference components is first set as a signal processing / interference cancellation processing target. In the following, since the layers to be processed are sequentially added in the order in which the accuracy of the replica is estimated to be high, signal processing / interference cancellation is performed. It is possible to generate a highly accurate replica and to obtain high signal detection accuracy. In addition, since the signal detection accuracy can be increased, it can be applied to communication with reduced transmission power.

  Embodiments of a receiving apparatus and a demodulation method according to the present invention will be described below in detail with reference to the drawings. Note that the present invention is not limited to the embodiments.

Embodiment.
FIG. 1 is a flowchart showing an example of a processing procedure of an embodiment of a demodulation method according to the present invention. In the present embodiment, an example using 2 ^2M QAM symbols will be described.

FIG. 2 is a diagram showing an example of symbol layering of 2 ^2M QAM symbols. As shown in FIG. 2, in the present embodiment, 2 ^2M QAM symbols are considered as symbols in which M QPSK symbols having different signal powers are hierarchized. That is, the QPSK symbols (sub-symbols) hierarchized from the 0th layer to the (M−1) th layer in descending order of amplitude are considered. The sub-symbol hierarchization is characterized in that the signal determination accuracy of the QPSK sub-symbol is lower and the replica generation accuracy is lower in the upper layer. In the present embodiment, signal detection / replica generation and interference cancellation processing are executed preferentially from lower QPSK symbols with high signal detection accuracy, so that signal detection / replica generation can be performed while maintaining the signal detection accuracy of the entire QAM symbol. Gradually expand the scope of the target audience. As a result, it is possible to effectively suppress the interference component while avoiding an increase in interference component removal error due to the low-precision replica that has occurred in the prior art.

  In addition, here, as an example, a case where a single carrier modulation scheme using frequency domain equalization is applied will be described, but the equalization scheme is not limited to frequency domain equalization, and the modulation scheme is not limited. It is not limited to single carrier modulation. In this example, a case where a multi-antenna is used will be described, but the present invention is not limited to the multi-antenna. Furthermore, it is applicable to optical communication and is not limited to wireless communication.

  In this embodiment, a QAM symbol is described as an example. However, the present invention is not limited to this, and multilevel PSK can be hierarchized as long as it is a symbol of 2 bits or more. In addition, when hierarchization can be performed so that the number of layers and the number of allocated bits after hierarchization satisfy the total number of bits of QAM or PSK, the number of bits can be arbitrarily set, and the amount of calculation and characteristics It is possible to design a hierarchical system flexibly from the viewpoint of

  Next, the demodulation method of this embodiment will be described with reference to FIG. First, equalization / interference cancellation processing is performed with the 0th layer being the lowest layer as a processing target (step S1). Next, an equalization / interference cancellation process for processing from the 0th layer to the first layer, an equalization / interference cancellation process for processing from the 0th layer to the second layer, and so on. And the processing is sequentially performed from the 0th layer to the equalization / interference cancellation processing for the (M−1) th layer as a processing target. Here, the processing contents of the processing targeted for processing from the 0th layer to the mth layer (0 <m <(M−1)) will be described in detail.

  Here, it is assumed that the amplitude is constant for each layer and the layer having a large amplitude has a high accuracy of replica, but when the amplitude for each layer is not constant, the 0th layer, Similarly, the layers to be processed may be sequentially increased from the 0th layer.

  Prior to the equalization / interference cancellation process for the 0th layer to the mth layer (0 <m <(M-1)) as the processing target, the equalization / interference cancellation process for the 0th layer (step S1) ) To signal detection / replica generation processing (step S2) for processing from the 0th layer to the (m−1) th layer.

Next, as signal detection / replica generation processing (step S3) for processing from the 0th layer to the mth layer, first, frequency domain equalization processing is performed as follows to obtain a reception signal after equalization ( Step S11). In the first frequency domain equalization process, the received signal is subjected to FFT (Fast Fourier Transform) processing and converted to a frequency domain signal. In this embodiment, signal detection / replica generation is repeatedly performed, and the number of repetitions is i _max . The received signal after equalization of the t-th transmitting antenna of the f-th frequency in the i (1 ≦ i ≦ i _max ) -th iterative process is obtained by detecting W (f) _t ^{i, m} in the 0th to ^m- th layers. When the t-th transmission antenna equalization weight vector is used as a replica generation target and R (f) is a received signal vector, it can be obtained based on the following equation (1).

W (f) _t ^{i, m} is (.) ^H is Hermitian transposition operation, (.) ⁻¹ is inverse matrix operation, S is transmission power, σ ² is noise power, and H (f) is transmission path. When a channel gain matrix obtained by estimation is used and G ^{i, m} is a matrix in which residual interference power estimation values are arranged, it can be obtained based on the following equation (2).

G ^{i, m} is a diagonal matrix having a matrix size corresponding to the number of transmission antennas, and the residual interference power estimation values of the transmission antennas corresponding to the respective diagonal elements are arranged. The residual interference power estimation value is a value corresponding to the interference power estimation value that cannot be reduced by the subsequent interference cancellation processing, and the 0th to mth layers generated in the previous iteration (i-1) signal detection / replica generation. Can be calculated from the sub-symbol replicas. In frequency domain equalization, only the residual interference power is suppressed, and the remaining degree of freedom is operated to acquire the antenna / frequency diversity gain.

Next, interference cancellation processing for subtracting the interference replica from the equalized reception signal obtained in step S11 is executed in the frequency domain (step S12). Specifically, for example, H _i bar (f) is an equivalent channel horizontal vector with the number of transmission antennas as the number of elements, and D ^{i, m} hat (f) is a replica vertical vector with the number of transmission antennas as the number of elements. At this time, the reception signal after the interference cancellation processing is obtained based on the following equation (3).

The replica vertical vector is obtained based on the sub-symbol replica in the frequency domain obtained by the FFT process in step S18 described later. In each element of the replica vertical vector D ^{i, m} hat (f), the 0th to mth layer sub-symbol replicas of each transmission antenna are multiplexed. t-th antenna of the D ^i, D _t ⁱ corresponding to ^m hat ^{(f), m} hat (f) ^{_{is, D t, m 'i,}} m hat (f) a signal detection / replica generation target zeroth When the frequency domain replica of the m'th layer of the i-th t-th transmitting antenna in the case of the m layer is used, it can be expressed by the following equation (4).

Next, IFFT (Inverse Fast Fourier Transform) processing is performed on the received signal after the interference cancellation processing (step S13), soft decision is performed on the processing target layer based on the received signal after IFFT, and the reliability Information is calculated and the calculated reliability is determined and fed back (step S14). Then, it is determined whether i is equal to i _max (step S15). If not equal (step S15 No), i is set to i + 1 (step S16), and the processing object is processed using the feedback reliability information. A sub-symbol replica of the layer is generated (step S17). The sub-symbol replica is subjected to FFT processing to obtain a frequency-domain sub-symbol replica (step S18), and the process returns to step S11 to repeat the processes after step S11. At this time, the sub-symbol replica calculated in step S18 is used for the interference cancellation process in step S12 described above. Also, the residual interference power estimation value used in step S11 is obtained based on the sub-symbol replica calculated in step S17 or S18.

If it is determined in step S15 that i is equal to i _max (step S15 Yes), the process proceeds to a process (step S4) in which the 0th to (m + 1) th layers are processed. Then, the processes similar to those in step S3 are performed by sequentially increasing the layers to be processed until the process (step S5) in which the 0th to (M-1) th layers are processed. In this way, the signal detection / interference cancellation process for all layers is finally completed.

  In the above description, the process when i is 2 or more has been described. However, since i = 1, that is, the frequency domain sub-symbol replica is not obtained in the first repetition process, step S12 is not performed. Then, the process proceeds from step S11 to step S13, and the processes after step S13 are performed on the received signal obtained in step S11. Further, the processes of steps S1, S2 and steps S4, 5 are the same as the process of step S3 except that the processing targets are different.

  In this embodiment, the soft decision output is used as reliability information for replica generation. However, the reliability information used for replica generation is not limited to this, and information obtained by other methods may be used. Good.

  In the prior art, the interference component could not be sufficiently reduced by using the low-accuracy sub-symbol replica, but by the above processing, the equalization / interference cancellation processing is sequentially performed from the layer with high sub-symbol replica generation accuracy. As a result, errors due to the use of low-precision sub-symbol replicas can be avoided, and high-accuracy signal detection becomes possible.

  FIG. 3 is a diagram illustrating a functional configuration example of a receiving apparatus that implements the demodulation method according to the present embodiment. As shown in FIG. 3, this receiving apparatus includes receiving antennas 1-1 to 1-N, analog signal processing units 2-1 to 2-N, A / D conversion units 3-1 to 3-N, and channel estimation processing. 4, a frequency domain equalization processing unit 5, an interference cancellation unit 6, a soft decision unit 7, a sub-symbol replica generation unit 8, a replica generation control unit 9, and an error correction decoder 10.

  The receiving operation of this receiving apparatus will be described. First, the N reception antennas 1-1 to 1-N receive the transmitted signals, and the analog signal processing units 2-1 to 2-N respectively perform predetermined analog reception processing on the corresponding reception signals. Do. Then, the A / D (Analog / Digital) conversion units 3-1 to 3-N convert the analog signals after reception processing into digital signals, respectively, to obtain time domain reception samples. The channel estimation unit 4 obtains a transmission path estimation value based on the time domain received samples output from the A / D conversion units 3-1 to 3-N.

  The frequency domain equalization processing unit 5 transmits from each transmission antenna based on the time domain reception samples output from the A / D conversion units 3-1 to 3-N and the channel estimation value obtained by the channel estimation processing unit 4. A frequency domain equalization process is performed on the transmitted signal. In the processing after the frequency domain equalization processing unit 5, as described above, first, the 0th layer is set as a processing target, and subsequently, the processing target layers are sequentially increased, and signal detection / interference is performed for the target layer. Repeat the cancellation process. In the first iterative process, the frequency domain equalization processing unit 5 performs an FFT process on the time domain received sample and performs a frequency domain equalization process after using the received signal in the frequency domain.

  In the first iterative process for the layer to be processed, the soft decision unit 7 performs an IFFT process on the received signal after the frequency domain equalization process, and then performs a soft decision process on the transmission signal of each transmitting antenna. Get reliability information. In the second and subsequent iterations, after the frequency domain equalization process, after the interference cancellation unit 6 performs the interference cancellation process of step S12 described above, the soft decision unit 7 performs a soft decision process on the received signal after the interference cancellation process. To obtain reliability information.

  The soft decision unit 7 outputs the obtained reliability information to the sub-symbol replica generation unit 8, and the sub-symbol replica generation unit 8 determines the sub-symbol replica of the processing target layer based on the instruction from the replica generation control unit 9. Is generated. Note that the replica generation control unit 9 sequentially instructs the sub-symbol replica generation unit 8 on the layer to be processed and manages the generated replica. That is, the replica generation control unit 9 can be considered as an overall control unit that manages the processing target layer of the signal detection / interference cancellation process. Then, the sub symbol replica generation unit 8 feeds back the generated sub symbol replica to the frequency domain equalization unit 5 and the interference cancellation unit 6.

  FIG. 4 is a diagram illustrating a functional configuration example of the frequency domain equalization unit 5. As shown in FIG. 4, the frequency domain equalization unit 5 includes a residual interference power calculation unit 51, an equalization weight calculation unit 52, and an equalization weight multiplication unit 53. The residual interference power calculation unit 51 estimates the residual interference power value using the sub-symbol replica, and the equalization weight calculation unit 52 calculates the equalization weight using the estimation result and the propagation path estimation value. And the equalization weight multiplication part 53 performs the frequency domain equalization demonstrated by the above-mentioned step S11.

  FIG. 5 is a diagram illustrating a functional configuration example of the interference cancellation unit 6. As shown in FIG. 5, the interference cancellation unit 6 includes an interference cancellation unit 61 and an interference replica generation unit 62. The interference replica generation unit 62 generates an interference replica that is a replica of the interference component of the received signal using the sub-symbol replica obtained in the previous iteration, the transmission path estimation value, and the equalization weight, The FFT processing is performed to obtain a frequency domain interference replica, and the interference cancellation unit 61 subtracts the frequency domain interference replica from the received signal after the frequency domain equalization processing (interference cancellation processing) as described in step S12 above. ).

Soft decision section 7 performs soft decision again using the received signal after the interference cancellation process, obtains reliability information, and outputs the reliability information to sub symbol replica generation section 8 again. Then, after repeating the frequency domain equalization processing a predetermined number of times (i _max described above), the number of layers to be processed is increased by one, and the same processing is performed. The same processing is performed until the processing target becomes the 0th to (M-1) th layers.

  In this embodiment, the sub-symbol replica is generated using the soft-decision output. However, the present invention is not limited to this, and the sub-symbol replica may be generated using the decoding result of the error correction code. FIG. 6 is a diagram illustrating a configuration example of a receiving device that generates a sub-symbol replica using a decoding result of an error correction code. As shown in FIG. 6, in this receiving apparatus, the frequency domain equalization processing unit 5 and the sub symbol replica generation unit 8 in FIG. 3 are replaced with an equalization processing unit 11 and a sub symbol replica generation unit 8a, respectively. Then, the sub symbol replica generation unit 8 a generates a sub symbol replica using the decoding result of the error correction decoder 10. The processing of the equalization processing unit 11 is the same as the processing of the frequency domain equalization unit 5. Other processing is the same as that of the receiving apparatus of FIG.

  Further, the generation of the sub-symbol replica using the soft decision output and the generation of the sub-symbol replica using the decoding result of the error correction code may be switched. FIG. 7 is a diagram illustrating a configuration example of a receiving apparatus that performs switching between generation of a sub-symbol replica using a soft decision output and generation of a sub-symbol replica using an error correction code decoding result. In the receiving apparatus shown in FIG. 7, the equalization processing unit 11 and the sub symbol replica generation unit 8a of the receiving apparatus in FIG. 6 are replaced with the equalization processing unit 11a and the sub symbol replica generation unit 8b, respectively, and a switch 12 is added. ing. By switching the switch 12, the input to the sub-symbol generator 8b can be switched between the soft decision output and the decoding result of turbo decoding.

  If the configuration as shown in FIG. 7 is adopted, for example, when adaptive modulation or the like that changes the coding rate is adopted, the reliability information from the error correction decoder 10 is fed back when the coding rate is low. When the coding rate is high, processing according to the coding rate, such as feeding back the soft decision output from the soft decision unit 7, can be performed. In addition, processing can be switched according to retransmission such as an automatic retransmission request. Further, when a quick response is required, the soft decision output from the soft decision unit 7 is fed back, and when a slight delay is allowed, the error correction decoder 10 feeds back the application request. It can also be switched based on the specification. Furthermore, it is possible to design a more flexible system by changing the control content instructed by such switching and the replica generation control unit 9.

  As described above, in the present embodiment, when multilevel modulation symbols are hierarchized into sub-symbols and demodulation is performed, signal processing / interference cancellation is first performed for a layer that is estimated to have the highest accuracy of interference component replicas. In the following, signal processing / interference cancellation is performed by adding layers to be processed in the order in which replica accuracy is estimated to be high. For this reason, it is possible to avoid an error due to the use of a low-precision sub-symbol replica, and it is possible to detect a signal with high accuracy. In addition, since the signal detection accuracy can be increased, it can be applied to communication with reduced transmission power, and energy saving as the entire communication system can be achieved.

  As described above, the receiving apparatus and the demodulation method according to the present invention are useful for a receiving apparatus that receives and demodulates a modulated signal, and are particularly suitable for a receiving apparatus that demodulates multilevel symbols as hierarchical subsymbols. .

It is a flowchart which shows an example of the process sequence of embodiment of the demodulation method concerning this invention. It is a figure which shows an example of the symbol hierarchization of a 2 ^2M QAM symbol. It is a figure which shows the function structural example of the receiver which implement | achieves the demodulation method of this invention. It is a figure which shows the function structural example of a frequency domain equalization part. It is a figure which shows the function structural example of an interference cancellation part. It is a figure which shows the structural example of the receiver which produces | generates a sub symbol replica using the decoding result of turbo decoding. It is a figure which shows the structural example of the receiver which switches and performs the production | generation of the subsymbol replica using a soft decision output, and the production | generation of a subsymbol replica using the decoding result of turbo decoding.

Explanation of symbols

1-1 to 1-N reception antenna 2-1 to 2-N analog signal processing unit 3-1 to 3-N A / D conversion unit 4 channel estimation processing unit 5 frequency domain equalization processing unit 6 interference cancellation unit 7 soft Determination unit 8, 8a, 8b Sub-symbol replica generation unit 9 Replica generation control unit 10 Error correction decoder 11, 11a Equalization processing unit

Claims (7)

A reception device that performs equalization processing and interference component removal processing for each layer of the hierarchies by regarding one symbol of a multilevel modulated reception signal as a plurality of sub-symbols,
A layer having the highest signal accuracy is designated as a processing target of each processing, and after the processing based on the instructions is completed, a layer having the next highest signal accuracy among the layers other than the processing target layer is provided to the processing target layer. The two layers are designated as the next processing target in descending order of signal accuracy, and the signal accuracy of the next processing target layer is subsequently increased until all layers in the hierarchization are processed. Control means for instructing the result of adding a higher layer as a processing target;
Sub-symbol replica generation means for generating a sub-symbol replica that is a replica of the sub-symbol of the layer to be processed based on the instruction;
Transmission path estimation means for performing transmission path estimation based on the sub-symbol replica;
With
A receiving apparatus, wherein the equalization process is performed based on an estimation result of the transmission path estimation unit, and the interference component removal process is performed based on the sub-symbol replica.   2. The signal accuracy is determined based on amplitude for each sub-symbol, SNR (Signal to noise power ratio) after reception, SINR (Signal to interference plus noise power ratio), or reliability information. The receiving device described in 1.   The receiving apparatus according to claim 1, wherein the equalization process and the interference component removal process are repeatedly executed a predetermined number of times for each processing target.   4. The receiving apparatus according to claim 1, wherein the sub-symbol replica generating unit further generates a sub-symbol replica based on reliability information obtained as a soft decision result of the received signal.   4. The receiving apparatus according to claim 1, wherein the sub-symbol replica generation unit further generates a sub-symbol replica based on reliability information obtained as an error correction decoding result of the received signal.   6. The receiving apparatus according to claim 1, wherein one symbol of the multilevel modulated reception signal is a QAM symbol, and the sub-symbol is a QPSK symbol. A demodulation method in the case where one symbol of a multilevel-modulated received signal is regarded as a plurality of sub-symbols that are hierarchized, and equalization processing and interference component removal processing are performed in units of the hierarchies,
A layer having the highest signal accuracy is designated as a processing target of each processing, and after the processing based on the instructions is completed, a layer having the next highest signal accuracy among the layers other than the processing target layer is provided to the processing target layer. The two layers are designated as the next processing target in descending order of signal accuracy, and the signal accuracy of the next processing target layer is subsequently increased until all layers in the hierarchization are processed. A control step for indicating the result of adding a higher layer as a processing target;
A sub-symbol replica generation step of generating a sub-symbol replica that is a replica of the sub-symbol of the layer to be processed based on the instruction;
A channel estimation step for performing channel estimation based on the sub-symbol replica;
An equalization processing step for performing the equalization processing based on the estimation result of the transmission path estimation;
An interference component removal step for performing the interference component removal processing based on the sub-symbol replica;
A demodulating method.

Images