JP2009130486A - Wireless communication system, and receiver - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a wireless communication system capable of decoding transmission data processed with error correction coding processing, in a high precision. <P>SOLUTION: The wireless communication system includes a transmitter for performing error correction coding processing to a signal and transmitting the signal and a receiver for performing error correction processing. The receiver estimates propagation path characteristics from a received pilot signal and estimates variance of noise in each modulated signal. The estimated propagation characteristics and the estimated variance of noise are used for decoding processing (maximum-likelihood sequence estimation and turbo decoding etc.). <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to a radio communication system and a receiving apparatus, and more particularly to a radio communication system and a receiving apparatus that can decode transmission data subjected to error correction coding processing with high accuracy.

In recent years, various wireless communication systems have been mixed due to the development of wireless communication technology.
In recent wireless communication systems, communication path encoding / decoding techniques are applied so that data can be demodulated not only at fluctuations in the wireless propagation path through which wireless signals pass, but also at lower reception levels.

  Among these channel encoding / decoding technologies, convolutional encoding and turbo encoding are particularly used recently, and decoding processing includes maximum likelihood sequence estimation (Viterbi decoding) and turbo decoding. Is already used in mobile phones and wireless local area networks (LANs).

  However, these are technologies originally developed in the field of coding theory, and only additive white Gaussian noise (AWGN) with uncorrelated Gaussian property in time or frequency is applied to the transmitted signal. In the case where the system is optimized in the case of giving a disturbance, it is possible to perform decoding using only received bits under the premise that the AWGN environment is used.

Here, maximum likelihood sequence estimation (MLSE) will be described. The equation used for determination in maximum likelihood sequence estimation is based on the assumption that the variance of noise in each signal is constant, where r (k) is the received signal at the kth sample and s (k) is the transmitted signal at that time. Is expressed as shown in Equation (1). In Equation (1), σ ² represents noise variance, and N represents a sequence length.

This equation (1) is based on the premise that each received signal is represented by a Gaussian distribution, and is represented by multiplication of the Gaussian distribution of each series. This represents the combined probability density function of the received signals of the entire sequence, and all possible transmission signals s (k) are brute-forced in all time series so that the value of equation (1) is maximized. This is a technique in which the sequence is a transmitted sequence.
Next, when Expression (1) is transformed, it is expressed as Expression (2).

Here, equation (2) takes the maximum value when the sum of | r (k) -s (k) | ² is minimized under the condition that the variance of all received signals is constant. I understand.
Therefore, when Viterbi decoding or the like for realizing MLSE is performed with only the received signal when the variance is constant, theoretically correct sequence estimation can be performed.

On the other hand, in the wireless communication system technology, since the wireless LAN is originally used as a hot spot, the disturbance is only the AWGN of the receiver and there is no problem.
However, in a cellular system typified by a recent mobile phone, a one-cell frequency repetition system that uses the same frequency in all cells is promising for the purpose of effective use of frequencies.

  In this case, although the decoding process regards the interference from the adjacent cell as the noise in the receiver and performs the decoding process, the interference is also received from the adjacent cell via the radio propagation path, so the receiver noise has no correlation. However, with respect to interference, a correlation may occur due to the impulse response of the propagation path.

Therefore, as a result, when interference and noise are regarded as noise at the receiver, the premise of AWGN is broken, and there arises a problem that transmission characteristics deteriorate near a cell boundary where interference is strong. In this case, since the variance σ ² in equation (1) depends on k, σ ² becomes σ ² (k), and equation (2) is expressed as equation (3).

In this case, it is necessary to calculate | r (k) −s (k) | ² / 2σ ² (k) (that is, Euclidean distance normalized by variance) and calculate a physical quantity that minimizes the sum. Don't be.

On the other hand, Patent Document 1 proposes a method of performing decoding processing as AWGN by estimating and normalizing the variance after calculating the Euclidean distance inside the decoder.
FIG. 11 is a diagram when the explanation of estimating and normalizing the variance value is simplified, although it is different from the configuration diagram of the receiver shown in Patent Document 1.

  In the figure, a receiver includes a pilot separation unit 1001, a propagation path estimation unit 1002, a reference signal generation unit 1003, a propagation path multiplication unit 1004, a Euclidean distance calculation unit 1005, a Euclidean distance variance estimation unit 1006, and a Euclidean distance normalization unit 1007 , A metric calculation unit 1008 and a surviving path selection unit 1009.

  First, the pilot signal for estimating the propagation path characteristic is separated from the received signal by the pilot separation section 1001, and the pilot signal is input to the propagation path estimation section 1002 to estimate the propagation path characteristic. At the same time, the received signal from which the pilot signal has been separated is input to Euclidean distance calculation section 1005.

  On the other hand, the reference signal generation unit 1003 generates a reference value of the generated transmission signal that can be transmitted, and the propagation path multiplication unit 1004 multiplies the propagation path characteristics to obtain a received signal that is not affected by noise or interference. Candidates are generated and input to the Euclidean calculation unit 1005.

  The Euclidean calculation unit 1005 calculates the Euclidean distance by subtracting the candidate of the received signal from the received signal from which the pilot signal is separated, and calculating the square of the absolute value. The Euclidean distance calculated in this way is estimated by the Euclidean distance variance estimation unit 1006 and normalized by the Euclidean distance normalization unit 1007.

  Thereafter, the metric calculation unit 1008 calculates the metric by adding the normalized Euclidean distance to the metric calculated up to one bit before, and the surviving path selection unit 1009 selects the surviving path.

Furthermore, this operation is repeated as many times as the number of transmission bits, and each bit is estimated by following a path in which the final state metric is minimized by a sequence estimation unit (not shown). Here, the reason why the two signals are output from the metric calculation unit 1008 is that there are two paths that transition to the same state.
JP 2006-286073 A

  However, the measurement of variance is obtained by averaging the square values of a plurality of signals having the same variance, and is essentially the same value to estimate after calculating the Euclidean distance for each received bit. It is necessary to measure the dispersion by extracting only the Euclidean distance of the transmission signal having such dispersion.

  Furthermore, since Patent Document 1 is intended to estimate a dispersion value that changes due to jitter of an optical disk, it is necessary to estimate in advance the dispersion that affects each bit in the decoder. .

  In addition, when this method is introduced into a wireless communication system, an interleaver or the like that randomly arranges bits is usually used in the wireless communication system. Therefore, even if the method disclosed in Patent Document 1 is introduced, the bit is affected. The received bits having substantially the same degree of dispersion are scattered, and an appropriate operation cannot be performed.

  The present invention has been made in consideration of the above-described circumstances, and an object thereof is to provide a radio communication system and a receiving apparatus that decode transmission data subjected to error correction coding processing with higher accuracy. To do.

In order to solve the above problems, a wireless communication system of the present invention includes a transmission device that performs error correction coding processing and transmits a signal, and a reception device that performs error correction processing,
The receiving apparatus includes a propagation path characteristic estimation unit that estimates a propagation path characteristic from a received pilot signal, an interference noise estimation unit that estimates a noise variance in each modulation signal, and the propagation path characteristic and the noise variance. A decoding unit that performs error correction processing on the received data signal.

Furthermore, the decoding unit uses the normalized Euclidean distance calculated using the noise variance when performing error correction processing.
The interference noise estimation unit estimates a variance of noise from the propagation path characteristic and the received pilot signal, or from the reception signal replica generated from the propagation path characteristic and data decoded by the decoding unit. The variance of noise was estimated.

  According to the present invention, for transmission data subjected to error correction coding processing, the cause of noise change in a wireless communication system is grasped, noise variance is estimated outside the decoding unit, and the variance is decoded. Therefore, decoding can be performed with higher accuracy.

Hereinafter, preferred embodiments of a wireless communication system and a receiving apparatus of the present invention will be described with reference to the drawings.
In the following embodiments, an orthogonal frequency division multiplexing (OFDM) method is used as a transmission method of a one-frequency repetitive cellular system, and interference from adjacent cells is subjected to frequency selective fading for each subcarrier. A case is assumed where the variance of noise (here, the sum of interference and noise in the receiving apparatus) is different.

  Even when the variance of noise changes with time, the variance can be estimated in the same way, and error correction processing can be performed in consideration of the variance. Therefore, it is possible to introduce even a single carrier method in which the variance changes for each symbol. .

<First Embodiment>
In the first embodiment, it is assumed that maximum likelihood sequence estimation is used as a decoding algorithm. Generally, a Viterbi algorithm is often used as an algorithm for realizing this maximum likelihood sequence estimation.

  FIG. 1 is a schematic diagram of a wireless communication system according to the first embodiment. In the figure, 10a is a first base station apparatus that transmits a desired signal, 10b is a second base station apparatus that is communicating with another terminal in an adjacent cell and transmits an interference signal, and 20a is a first base station apparatus. It is a terminal device communicating with the base station device 10a.

In this case, for the terminal device 20a, the signal 30a transmitted from the first base station device 10a is a desired signal, and the signal 40b transmitted from the second base station device 10b is an interference signal.
For the terminal device 20b, the signal 30b transmitted from the second base station device 10b is a desired signal, and the signal 40a transmitted from the first base station device 10a is an interference signal.

  Here, when the OFDM scheme is assumed, the frequency characteristics of the desired signal and the frequency characteristics of the interference signal and the thermal noise in the receiving apparatus determine the reception status of each subcarrier. Therefore, the sum of interference and noise of each subcarrier corresponding to noise has frequency selectivity, that is, correlation, and as a result, noise dispersion varies among subcarriers. Therefore, Equation (3) is maximized. It is necessary to estimate such a sequence.

  FIG. 2 is an example of a transmission apparatus according to the first embodiment. In the figure, a transmission apparatus includes an encoding unit 100, an interleaver unit 101, a modulation unit 102, an S / P (serial-parallel) conversion unit 103, an IDFT (Inverse Discrete Fourier Transform) unit 104, a P / S. (Parallel to serial) conversion unit 105, pilot signal generation unit 106, pilot multiplexing unit 107, GI (Guard Interval) insertion unit 108, radio unit 109, and transmission antenna 110 are included.

  First, transmission bits are channel-coded by the encoding unit 100, the code bits are rearranged by the interleaver unit 101, and modulated by the modulation unit 102. Thereafter, the modulation data is parallelized by the S / P conversion unit 103 and converted into a time signal by the IDFT unit 104. The time signal is serialized by the P / S converter 105.

On the other hand, the propagation path estimation signal generated by the pilot signal generation unit 106 is multiplexed with the serialized time signal by the pilot multiplexing unit 107, and a guard interval is inserted by the GI insertion unit 108.
Finally, it is up-converted to a radio frequency by the radio unit 109 and transmitted by the transmission antenna 110.

  FIG. 3 is an example of a receiving apparatus according to the first embodiment. In the figure, the receiving apparatus includes a receiving antenna 111, a radio unit 112, a GI removing unit 113, a pilot separating unit 114, a propagation path characteristic estimating unit 115, an interference noise estimating unit 116, an S / P converting unit 117, a DFT unit 118, The P / S conversion unit 119, the reference signal generation unit 120, the propagation path characteristic multiplication unit 121, the Euclidean distance calculation unit 122, the Euclidean distance normalization unit 123, the deinterleaver unit 124, and the maximum likelihood sequence estimation unit 125 are configured.

  Here, the reference signal generation unit 120, the propagation path characteristic multiplication unit 121, the Euclidean distance calculation unit 122, the Euclidean distance normalization unit 123, the deinterleaver unit 124, and the maximum likelihood sequence estimation unit 125 are generally referred to as decoding. The received signal, noise variance, and propagation path characteristics are input to the decoding unit.

First, the received signal is received by the receiving antenna 111 and is down-converted from the radio frequency by the radio unit 112.
Next, the guard interval is removed by the GI removal unit 113 and the pilot signal is separated into the pilot signal and the data signal by the pilot separation unit 114.

The propagation path characteristic estimation unit 115 estimates the frequency response of the propagation path through which the desired signal has passed from the separated pilot signal. Here, the frequency response of the propagation path through which the desired signal has passed is obtained by applying a time window corresponding to the maximum delay time of the delay wave designed according to the cell radius etc. to the impulse response of the propagation path, and thereby including noise including interference. It is estimated by suppressing.
Interference noise estimation section 116 estimates the noise variance of each subcarrier using the channel frequency response thus estimated and the original received pilot signal.

  On the other hand, the separated data signal is parallelized by the S / P converter 117 and then frequency-converted by the DFT unit 118 to separate each subcarrier. Each subcarrier signal is serialized by the P / S converter 119.

Next, processing of the decoding unit will be described.
First, the reference signal generation unit 120 generates candidates for all transmission signals that can be transmitted, and then the propagation path characteristic multiplication unit 121 uses the estimated frequency response of the propagation path to receive all received signals. Then, the Euclidean distance calculation unit 122 calculates the Euclidean distance by calculating the distance between the received signal and the received signal candidate.

  Next, the Euclidean distance normalization unit 123 calculates a normalized Euclidean distance from the noise variance estimated by the interference noise estimation unit 116 and the Euclidean distance calculated by the Euclidean distance calculation unit 122.

  In the case of binary phase shift keying (BPSK), 1-bit code bit information is included in 1 subcarrier, so the subcarrier and the bit index are the same, so the Euclidean distance of the kth bit. Is calculated by equation (4).

In equation (4), d (k, 1) and d (k, 0) are normalized Euclidean distances in the kth subcarrier when the transmission bits are 1 and 0, respectively, and R (k) is k. The received signal of the th subcarrier and the output of the P / S conversion section 119, and H _est (k) is the complex gain of the propagation path in the kth subcarrier and the output of the propagation path characteristic estimation section 115. R (k) and H _est (k) are represented by complex numbers. S _ref (k, 1) and S _ref (k, 0) are transmission signal candidates when the transmission bit is 1 and 0, and are the outputs of the reference signal generation unit 120. Note that σ ² (k) is an output of the interference noise estimation unit 116, and a typical calculation method of this value, that is, a detailed operation of the interference noise estimation unit 116 will be described later. Also, σ ² (k) is calculated for each OFDM symbol.

The propagation path characteristic multiplication unit 121 calculates H _est (k) × S _ref (k, m) in Expression (4). Then, the Euclidean distance calculation unit 122 performs an operation corresponding to the numerator of Expression (4). Further, the Euclidean distance normalization unit 123 performs division shown in the denominator of Expression (4), and outputs d (k, m).

Further, in the expression (4), the case of BPSK has been described, but even in the case of other modulation schemes, it can be used for calculating the Euclidean distance in each code bit.
For example, in the case of QPSK (Quadrature Phase Shift Keying) in which information of 2 bits of code bits is carried in one subcarrier, the following equation (4 ′) is divided into the first and second bits constituting the QPSK symbol: It is calculated by the equation (4 ″).

In the equations (4 ′) and (4 ″), the values of the first and second bits constituting the QPSK symbol, that is, the value of 0 or 1, are substituted for m ₁ and m ₂ , respectively.
For example, when calculating the Euclidean distance when the first bit constituting the QPSK symbol is m ₁ = 0, out of the two signal points S _ref (k, 0, m ₂ ) of QPSK where the first bit is 0 The one having the smallest value is defined as the Euclidean distance in the equation (4 ′). By calculating in this way, calculation is possible even if a modulation method including several bits in one subcarrier is adopted.

  Thereafter, the normalized Euclidean distance that is the output of the Euclidean distance normalization unit 123 returns the bit rearrangement to the original by the deinterleaver unit 124, and finally the maximum likelihood sequence estimation unit 125 transmits the transmission with the highest probability. A bit sequence is estimated.

Next, an example of a method for estimating the frequency response of the propagation path and a method for estimating the noise will be described.
Here, a case where both propagation path characteristic estimation and noise estimation are performed using pilot signals, that is, the above-described propagation path characteristic estimation unit 115 and interference noise estimation unit 116 will be described.

First, the propagation path characteristic estimation unit 115 will be described.
FIG. 4 is an example showing a configuration of the propagation path characteristic estimation unit 115. In the figure, the propagation path characteristic estimation unit 115 includes a first DFT unit 201, a complex division unit 202, an IDFT unit 203, a time filter unit 204, and a second DFT unit 205.

  The received pilot signal is converted into a frequency axis signal by the first DFT unit 201, and the pilot signal becomes a known signal on the transmission / reception side. Get a response once.

In order to further reduce the influence of noise from the obtained frequency response, the IDFT unit 203 converts it into an impulse response of the propagation path on the time axis. Here, since the impulse response of the propagation path is only the length of the maximum delay time of the delay wave in the wireless propagation path, the time filter unit 204 suppresses or eliminates noise including interference on samples after the maximum delay time. To do.
Thereafter, the second DFT unit 205 converts the impulse response in which the influence of noise is reduced into a frequency response, thereby obtaining a frequency response of the propagation path.

Here, it is known that the estimation accuracy of the propagation path in this case is greatly improved by the time filter. In general, the length of the impulse response is designed to be 10 to 20% of the OFDM symbol in the case of OFDM. Since the effective symbol length for actually detecting data is 80% of the OFDM symbol length in the case of 20%, the effective symbol interval / guard interval interval = 80/20 = 4, that is, decibels. When converted, a filter gain of 10 × log ₁₀ 4 = 6 dB is obtained.

  Thus, since the propagation path estimation accuracy can be increased from the signal-to-noise ratio at the time of reception by the time filter, the interference noise power can also be estimated using the estimated frequency response of the propagation path.

Next, the interference noise estimation unit 116 will be described.
FIG. 5 is an example showing the configuration of the interference noise estimation unit 116. In the figure, the interference noise estimation unit 116 includes a DFT unit 211, a complex multiplication unit 212, an error measurement unit 213, and an averaging unit 214.

First, the DFT unit 211 converts the received pilot signal into a frequency axis received signal. On the other hand, the complex multiplier 212 multiplies the transmission pilot signal by the estimated channel frequency response to obtain a reference pilot signal that is not affected by interference or noise.
In error measurement section 213, a square error due to noise including interference is obtained from equation (9) from the obtained reference pilot signal and received pilot signal.

In the above formula (9), e (k) is the square error, R p _(k) is a received pilot signal of the k-th subcarrier, H _{est (k)} is the channel estimation unit of the k-th subcarrier The complex gain S _p (k) of the propagation path in the k-th subcarrier estimated at 115 is a known transmission pilot signal.

  The reception pilot signal is obtained by the above-described pilot separation unit 114. Usually, the transmission pilot signal is a known signal on the transmission / reception side and is transmitted to measure how much the transmission pilot signal is distorted. Therefore, the transmission pilot signal is known on the receiving side and is used.

Next, the averaging unit 214 averages the square errors obtained in this way.
In the case of OFDM, since the length of the impulse response of the propagation path is shorter than the OFDM symbol length, the variation in interference between close frequencies is correlated on the frequency axis. Therefore, when averaging within an OFDM symbol, a plurality of adjacent subcarriers are blocked, and the average value within the block is treated as the variance of noise within the block.

Further, when a plurality of OFDM symbols are time-multiplexed into frames, averaging may be performed using the same subcarrier number within a range in which the time variation of the propagation path is moderate.
This average value can be calculated by Equation (10), where X is the number of OFDM symbols averaged in the time and frequency directions, and Y is the number of subcarriers in the block for averaging in the frequency direction.

  In the above equation (10), e (x, y) is a square error calculated by equation (9) in the y-th subcarrier of the x-th OFDM symbol in the frame. K is all subcarrier indexes included in the range of subcarriers from y ′ to y ′ + Y.

  Thus, in a wireless communication system, noise dispersion often depends on the propagation path, and noise is estimated by the interference noise estimation unit 116 using this feature, and the value is input to the decoding unit, which is high. Transmission data can be detected with reliability.

  In general, the correlation of noise including interference depends on the propagation path, and the correlation increases between adjacent subcarriers. Therefore, when variance is estimated, averaging between adjacent subcarriers is required. The estimation accuracy of variance is higher than that calculated in the decoding unit.

  Furthermore, when bits are rearranged in the transmission device, such as interleaving, the correlation of noise dispersion on each bit becomes lower between adjacent bits, making it difficult to estimate the dispersion. Since the correlation of the propagation path is used instead of the arrangement of the interleaves, there is an effect that the estimation of the dispersion is not related to the presence or absence of interleaving.

  This estimation method is also used for noise estimation required when equalization such as frequency domain equalization (FDE) is used in the DFT-S-OFDM (DFT-Spread-OFDM) method based on the OFDM method. Applicable.

Next, effects of the receiving apparatus according to the first embodiment will be described.
FIG. 6 is an example illustrating a result of using the receiving apparatus according to the first embodiment. In the figure, the horizontal axis represents received E _s / N ₀ , and the vertical axis represents the bit error rate. Here, E _s denotes an average noise power density energy, N ₀ is, including the interference per subcarrier.

  The solid curve in the figure represents the characteristic of calculating the metric by normalizing the Euclidean distance considering the variance value, and the dotted curve using only the Euclidean distance without considering the noise variance value. The decrypted characteristics are shown. Not considering variance is equivalent to setting the denominator in equation (4) to 1 (of course, it is not necessary to be 1 if it is a constant). In FIG. 3, the Euclidean distance normalization unit This corresponds to the case where 123 is not provided. From FIG. 6, it can be confirmed that the receiving apparatus of the first embodiment is effective.

<Second Embodiment>
In the second embodiment, the decoding algorithm used in the first embodiment is replaced with a maximum a posteriori (MAP) estimation algorithm.
In general, in the MAP algorithm, a BCJR (Bahl-Cocke Jelinek-Raviv) algorithm is often used, and it is assumed that this method is used.

  FIG. 7 is an example of a receiving apparatus according to the second embodiment. In the figure, the receiving apparatus includes a receiving antenna 111, a radio unit 112, a GI removing unit 113, a pilot separating unit 114, a propagation path characteristic estimating unit 115, an interference noise estimating unit 116, an S / P converting unit 117, a DFT unit 118, The P / S conversion unit 119, the reference signal generation unit 120, the propagation path characteristic multiplication unit 121, the Euclidean distance calculation unit 122, the Euclidean distance normalization unit 123, the deinterleaver unit 124, and the MAP estimation unit 126 are configured.

Similarly to the first embodiment, the block includes a reference signal generation unit 120, a propagation path characteristic multiplication unit 121, a Euclidean distance calculation unit 122, a Euclidean distance normalization unit 123, a deinterleaver unit 124, and a MAP estimation unit 126. Constitutes a MAP decoding unit.
In the figure, the reception processing up to the propagation path characteristic multiplication unit 121 is the same as that in FIG. The only difference from the first embodiment is the decoding process of the MAP estimation unit 126.

  In this way, in order to normalize the Euclidean distance, the noise variance is estimated from the received signal in advance, so that the reliability of the probability density function expressed by Equation (3) is improved even in MAP estimation, and decoding is performed. Performance is improved.

<Modification of Second Embodiment>
In general, turbo decoding calculates the likelihood of each bit from the output of one decoding unit, passes it to the other decoding unit, and repeats this series of processes a predetermined number of times to gradually improve the reliability of the transmitted bits. It is a technique to improve.
Therefore, turbo decoding can be applied by applying the iterative process using two decoding units described in the second embodiment.

  FIG. 8 is an example of a receiving apparatus when turbo decoding is used. In the figure, the receiving apparatus includes a receiving antenna 111, a radio unit 112, a GI removing unit 113, a pilot separating unit 114, a propagation path characteristic estimating unit 115, an interference noise estimating unit 116, an S / P converting unit 117, a DFT unit 118, P / S conversion section 119, reference signal generation section 120, propagation path characteristic multiplication section 121, Euclidean distance calculation section 122, Euclidean distance normalization section 123, first MAP estimation section 137, first prior information subtraction section 128, The first interleaver unit 129, the second interleaver unit 130, the second MAP estimation unit 131, the deinterleaver unit 132, and the second prior information subtraction unit 133 are configured. However, it is assumed that turbo coding is adopted for the encoding unit of the transmission apparatus at this time.

Similarly to the second embodiment, the reference signal generation unit 120, the propagation path characteristic multiplication unit 121, the Euclidean distance calculation unit 122, the Euclidean distance normalization unit 123, the first MAP estimation unit 137, and the first prior information A block comprising a subtractor 128, a first interleaver 129, a second interleaver 130, a second MAP estimator 131, a deinterleaver 132, and a second prior information subtractor 133 is a turbo decoder. It is composed.
In FIG. 8, the same reference numerals are given to components representing the same functions as those in FIG. 7, and differences will be described.

  The normalized Euclidean distance is first input to the first MAP estimation unit 137, and after error correction processing is performed, reliability information obtained by the error correction processing is input to the first prior information subtraction unit 128. Is done. Here, the prior information is a physical quantity representing the reliability of whether the input bit improved until the previous error correction process performed in the iteration is 0 or 1. Therefore, no subtraction is performed at the first time.

Next, the prior information of each bit obtained from the first prior information subtraction unit 128 is rearranged by the first interleaver unit 129.
In the second MAP estimation unit 131, the Euclidean distance obtained by the Euclidean distance normalization unit 123 is rearranged by the second interleaver unit 130, and the prior arrangement is performed by the first interleaver unit 129. By inputting information and performing error correction processing, the reliability of each transmission bit is obtained with higher reliability than the previous MAP estimation.
The obtained reliability is returned to the original order by the deinterleaver unit 132, and the prior information input by the second prior information subtraction unit 133 is subtracted.

  On the other hand, the second prior information subtraction unit 133 improves the second MAP estimation unit 131 by subtracting the input reliability from the bit reliability output from the second MAP estimation unit 131. The reliability is input as prior information of the first MAP estimation unit 137.

  By repeating the above processing a predetermined number of times, turbo decoding becomes possible. In this case as well, the reliability of each transmission bit can be improved by the Euclidean distance normalization unit 123, and the characteristics can be improved.

<Third Embodiment>
FIG. 9 shows an example applicable when framing is performed by temporally multiplexing a plurality of OFDM signals. This corresponds to the case where the variance estimated from the decoded data sequence is used. In frame A, decoding is performed using the noise variance estimated from the previous frame, and then decoding is performed. The replica of the transmission signal is generated again using the result, and the variance of noise in each subcarrier is estimated by calculating the distance from the reception signal.

  FIG. 10 is an example of a receiving device according to the third embodiment. In the figure, the receiving apparatus includes a receiving antenna 111, a radio unit 112, a GI removing unit 113, a pilot separating unit 114, a propagation path characteristic estimating unit 115, an S / P converting unit 117, a DFT unit 118, and a P / S converting unit 119. , A reference signal generation unit 120, a propagation path characteristic multiplication unit 121, an Euclidean distance calculation unit 122, an Euclidean distance normalization unit 123, a memory 134, a sequence estimation unit 136, a signal replica generation unit 137, and an interference noise estimation unit 116. . In the same figure, the components representing the same functions as those in FIG.

The reception signal received by the reception antenna 111 is down-converted from the radio frequency by the radio unit 112, and the guard interval is removed by the GI removal unit 113. The signal from which the guard interval has been removed is separated into a pilot signal for propagation path estimation and a data signal, which is a signal carrying transmission data, in pilot separation section 114.
The separated pilot signal is input to the propagation path characteristic estimation unit 115, and the frequency response of the propagation path is calculated.

  On the other hand, the separated data signal is parallelized by the S / P converter 117 and converted into a signal of each subcarrier by the DFT unit 118. Thereafter, serialization is performed by the P / S converter 119.

Further, the reference signal generation unit 120 generates transmission signal candidates for all possible subcarriers, and the propagation path characteristic multiplication unit 121 multiplies the frequency response of the propagation path to calculate reception signal candidates.
Thereafter, the Euclidean distance is calculated in the Euclidean distance calculation unit 122 using the received signal of each subcarrier and the candidate for the received signal.

  Next, the memory 134 stores the noise variance of each subcarrier estimated in the previous frame. The Euclidean distance normalization unit 123 stores the Euclidean distance of each subcarrier stored in the memory 134. Normalize by noise variance.

  Next, maximum likelihood sequence estimation or MAP estimation is performed in sequence estimation section 136 for the normalized Euclidean distance, and the transmission bit is determined.

  Further, the signal replica generation unit 137 generates a replica of each subcarrier from the determined decoded data, and generates a reception signal replica obtained from the frequency response of the propagation path estimated as the transmission signal replica.

  Thereafter, the interference noise estimation unit 116 estimates the noise variance of each subcarrier from the serially received signal of each subcarrier and the received signal replica, and stores it in the memory 134 to store it as the noise variance of the next frame. To do.

  When demodulating the next frame, the stored noise value is used to detect the transmission bit of the next frame, that is, as an alternative to the interference noise estimation unit 116 of the first and second embodiments described above. be able to.

  Note that the present invention is not limited to the above-described embodiment, and various modifications and corrections can be made without departing from the scope of the present invention.

It is a schematic diagram of the radio | wireless communications system in 1st Embodiment. It is an example of the transmitter which concerns on 1st Embodiment. It is an example of the receiver which concerns on 1st Embodiment. It is an example which shows the structure of a propagation path characteristic estimation part. It is an example which shows the structure of an interference noise estimation part. It is an example which shows the result using the receiver of 1st Embodiment. It is an example of the receiver which concerns on 2nd Embodiment. It is an example of the receiver which concerns on the modification (when turbo decoding is used) of 2nd Embodiment. An example applicable when framing is performed by temporally multiplexing a plurality of OFDM signals is shown. It is an example of the receiver which concerns on 3rd Embodiment. It is a block diagram of the receiver in a prior art.

Explanation of symbols

DESCRIPTION OF SYMBOLS 10a ... 1st base station apparatus, 10b ... 2nd base station apparatus, 20a, 20b ... Terminal device, 30a, 30b ... Desired signal, 40a, 40b ... Interference signal, 100 ... Encoding part, 101 ... Interleaver part, 102 ... Modulation unit 103 ... S / P conversion unit 104 ... IDFT unit 105 ... P / S conversion unit 106 ... Pilot signal generation unit 107 ... Pilot multiplexing unit 108 ... GI insertion unit 109 ... Radio unit 110 ... Transmitting antenna, 111 ... receiving antenna, 112 ... radio unit, 113 ... GI removal unit, 114 ... pilot separation unit, 115 ... propagation channel estimation unit, 116 ... interference noise estimation unit, 117 ... S / P conversion unit, 118 ... DFT unit, 119 ... P / S conversion unit, 120 ... reference signal generation unit, 121 ... propagation path characteristic multiplication unit, 122 ... Euclidean distance calculation unit, 123 ... Euclidean distance positive , 124 ... deinterleaver, 125 ... maximum likelihood sequence estimator, 126 ... MAP estimator, 127 ... first MAP estimator, 128 ... first prior information subtractor, 129 ... first interleaver , 130 ... second interleaver unit, 131 ... second MAP estimation unit, 132 ... deinterleaver unit, 133 ... second prior information subtraction unit, 134 ... memory, 136 ... sequence estimation unit, 137 ... signal Replica generating unit 201 ... first DFT unit 202 ... complex division unit 203 ... IDFT unit 204 ... time filter unit 205 ... second DFT unit 211 ... DFT unit 212 ... complex multiplier unit 213 ... Error measurement unit, 214 ... averaging unit, 1001 ... pilot separation unit, 1002 ... propagation path estimation unit, 1003 ... reference signal generation unit, 1004 ... propagation path multiplication unit, 1005 ... Euclidean Away calculating unit, 1005 ... Euclidean calculation unit, 1006 ... Euclidean distance variance estimation unit, 1007 ... Euclidean distance normalization unit, 1008 ... metric calculation unit, 1009 ... survivor path selection unit.

Claims (6)

  A receiving apparatus that communicates with a transmitting apparatus that performs error correction coding processing and transmits a signal, and that performs propagation characteristic estimation from a received pilot signal, and dispersion of noise in each modulated signal. An interference noise estimation unit for estimation, and a decoding unit that performs error correction processing on a data signal received from the propagation path characteristics and the noise variance.   The receiving apparatus according to claim 1, wherein the decoding unit normalizes a Euclidean distance using the variance of the noise.   The receiving apparatus according to claim 1, wherein the interference noise estimation unit estimates noise variance from the propagation path characteristics and the received pilot signal.   3. The reception apparatus according to claim 1, wherein the interference noise estimation unit estimates a variance of noise from the propagation path characteristics and a received signal replica generated from data decoded by the decoding unit. A receiving device.   In a wireless communication system including a transmission device that performs error correction coding processing and transmits a signal, and a reception device that performs error correction processing, the reception device estimates propagation path characteristics from the received pilot signal, and each modulation A radio communication system characterized by estimating noise variance in a signal and using the noise variance for decoding processing.   6. The wireless communication system according to claim 5, wherein the decoding process normalizes a Euclidean distance calculated from a received signal by the variance of the noise.

Images