JPH0514213A - Soft decision decoding method - Google Patents

Abstract

PURPOSE:To operate a high accuracy decoding by reducing the bit error rate of an original signal. CONSTITUTION:A phase and a distance from an origin are computed by phase and distance arithmetic processings S11 and S14, from coordinate information (b) on a phase space obtained by demodulating a signal modulated with a PSK modulation system, and the phase is transmitted to a phase selecting processing S12. The inputted phase is compared with the specific transmitted by the PSK modulation system, and the inherent specific phase whose phase shift has a small absolute value is selected, by the phase selecting processing S12. A phase likelihood is operated based on the phase shift by a phase likelihood arithmetic processing S13, and a corrected phase likelihood is searched based on the arithmetic result by a phase likelihood correcting processing S16. This corrected phase likelihood is used for a soft decision decoding.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a soft decision decoding method in wireless communication such as cellular mobile communication which adopts a phase modulation method such as phase shift keying (PSK).

[0002]

2. Description of the Related Art Conventionally, as a technique in such a field,
For example, some documents were described in the following documents.

Reference 1: Murano and Kaijo, "Digital Signal Processing in Information and Communications," First Edition (Sho 62-11-25), Shokoido P. 32-91 Reference 2; IE Transactions on Communications Technology (IEEE T
ransactions On Communications Technology), COM
-19 [5] (1971-10) (US) A. J. VIT
ERBI “Convolutional Codes and Their Performance in
Communications Systems) "P.751-772, Reference 3; B.SKLAR" DIGITAL COMMUNICATIONS
S) "(1988) PRENICE HALL (US)
sec. 6.3.4, P.I. 333-337 Conventionally, there is one described in Document 1 regarding the decoding of the phase modulation system, and PS that allocates a bit string to the phase of the carrier wave
The K modulation method will be described.

Considering that θi is the phase of the carrier wave and fc is the modulation frequency, and the phase θi changes at each symbol transmission interval T, the modulated wave is given by the following equation (1). _∞ h (t) is a roll-off filter H set so that the impulse response does not interfere with adjacent symbols.
It is a time domain function of (f). This h (t) satisfies the following conditions. h (t) = 1 (t = 0) h (t) = 0 (t = nT; t ≠ 0) m-phase PSK uses m θi. For example, in the case of 4-phase PSK (QPSK), 2 bits are assigned to one phase, and when phase 0 (0, 0), phase π / 2 (0, 1), and phase π (1 , 1), and (1, 0) is transmitted when the phase is 3π / 2. From equation (1), it can be seen that PSK modulation can be obtained by adding sin (θi) and cos (θi) respectively modulated with orthogonal carrier waves of frequency fc.

On the receiving side, a hessband signal is obtained by multiplying equation (1) by an orthogonal carrier. S in equation (1)
When (t) is multiplied by the orthogonal carrier wave exp (j2π fc t), the equation (2) is obtained.

[0006]

[Equation 1]

By removing the harmonic component of the equation (2), the baseband signal wi can be obtained.

However, in the actual communication, it is necessary to consider the frequency characteristics of the transmission path, the phase fluctuation with respect to the carrier wave, the influence of the frequency offset and the like. Hereinafter, with reference to FIG. 2, processing in the receiving unit of the conventional decoding method will be described.

FIG. 2 is a block diagram showing the configuration of a receiving section of a conventional radio signal transmitting / receiving apparatus.

In this receiving section, a demodulation section 10 for demodulating a radio frequency band signal (RF band reception signal) a affected by multipath fading in a radio line and detecting coordinate information b in a phase space, A hard-decision decoding unit 20 that corrects a bit error of the baseband signal based on the coordinate information b to obtain a reproduction signal c.

The demodulator 10 includes a bandpass filter 11,
Analog / digital converter (A / D converter) 12, automatic gain controller (AGC) 13, quadrature carrier generator 14,
The low pass filters 15a and 15b, the equalizer 16 and the phase synchronization circuit 17 are included. Further, the hard decision decoding unit 20 includes a data decision unit 21, a deinterleave unit 2
2 and the decoding unit 23.

In the processing of this receiving section, the modulated wave of the above formula (1) is affected by the frequency characteristics of the transmission line, phase fluctuation (jitter) with respect to the carrier wave, frequency deviation (offset), etc. The received waveform of a (= r (t)) is expressed as in equation (3). Here, go (t); product of time domain function g (t) of frequency characteristic G (f) of the transmission path and exp (j2π fc t) ga; fluctuation of level due to loss in the transmission path fo; frequency Offset θ (t); Phase Jitter The RF band received signal a of the equation (3) has a predetermined frequency band extracted by the band pass filter 11, and the A / D converter 12
Is converted into a digital signal and becomes a discrete signal. (3)
Since the gain constant ga in the expression varies depending on the state of the transmission line, the AGC 13 corrects the received signal level so that the received signal level is kept constant so that the subsequent digital signal processing can be performed at the same level. ..

In equation (2), the orthogonal carrier wave is expressed as exp (j2
π fc t), but actually, the received wave is divided into two
Cos (2π fc t) and sin (2π
fct) is multiplied by the value at the sampling point. The orthogonal carrier wave is generated by the orthogonal carrier wave generator 14. The low-pass filters 15a and 15b remove harmonic components from each of the received waves multiplied by the orthogonal carrier waves.

When the influence of the frequency characteristic of the transmission path or the phase fluctuation or frequency shift on the carrier is small, the baseband signal can be used as it is as the coordinate information b. When the influence is large, the equalizer 16 causes the frequency characteristic go (n-
Correct k). Further, the phase synchronization circuit 17 causes the phase variation 2πn fo / fs with respect to the carrier wave and the frequency shift θ.
Correct (n).

The signal point arrangement on the two-dimensional plane of the received complex baseband signal thus obtained is the coordinate information b. The coordinate information b does not accurately take the transmission phase θi because the correction of the influence of the frequency characteristic of the transmission path or the phase variation with respect to the carrier wave, the frequency shift, etc. is not necessarily appropriate. Therefore, the data decision unit 21 in the hard decision decoding unit 20 predicts the transmission symbol by selecting the transmission phase closest to the coordinate information b. The determination of the transmission symbol is uniquely determined depending on which region the coordinate information b falls in based on the determination boundary set on the two-dimensional plane. The data determined uniquely is rearranged by the deinterleave unit 22 and then subjected to a decoding process by the decoding unit 23.

The interleaved conversion is a conversion in which the signal bits input to the memory are rearranged and output, and has the effect of replacing consecutively generated bit errors with random errors. In this way, the hard-decision decoding uses the bit string corresponding to the estimated symbol as it is in the decoding process.
The decoding unit 23 performs a decoding process according to the encoding of the block code, the convolutional code and the like. For example, Viterbi decoding is performed on the convolutional code.

Next, the convolutional code and the hard decision and the soft decision in the Viterbi decoding described in the documents 2 and 3 will be described.

Generally, in wireless communication such as mobile communication and satellite communication, various measures such as diversity reception, equalization, and code error control are taken in order to improve the quality deterioration of signals in a wireless line. Convolutional coding, which is a type of code error control, is performed based on a convolutional code generation rule uniquely determined by a coding rate, a constraint length, and a generator polynomial. A graphical representation of this generation rule is a kind of state transition diagram called a trellis figure. The fold-in code is
At the time of decoding, it is possible to correct the bit error of the received signal by comparing the received signal with possible paths on the trellis pattern and selecting the most likely path (optimal path). Viterbi decoding is the most general method of decoding convolutional codes. It uses hard decision to compare with a selectable signal sequence of trellis figure by the signal value itself and the probability that the signal value takes that value (likelihood). There is a soft decision to compare with.

The Viterbi decoding method is described in the above reference 2, and the method will be described with reference to FIGS. 3 and 4.

FIG. 3 is an explanatory diagram of convolutional coding.

When convolutional coding is performed, the coding rate becomes m / n when output n bits are generated for input m bits. When the output is generated from the past k bits including the latest input bit, it is called the constraint length k. In this case, n generator polynomials of length k are required. FIG. 3 shows the coding rate 1/2, the constraint length 3, the generator polynomials 111 and 101.
Shows the case.

In FIG. 3, 3 bits including the latest input bit are stored in the buffer 25, and by convolution, a 2-bit output is obtained. Since the generator polynomials are 111 and 101, one of the outputs is the logical sum of all the bits of the buffer 25, and the other is the logical sum of the first and third bits of the buffer 25.

FIG. 4 is a trellis diagram which is a state transition diagram of the convolutional coding generation rule of FIG.

The vertical direction of FIG. 4 shows the state in the buffer 25 that does not include the latest bit, and a state of 2 ^{k- 1} occurs.
In the example, it is 4. In each state, when 0 is input, the state moves to the next state along the solid line, and 2 bits on the line are output. When 1 is input, the state moves to the next state along the broken line, and 2 bits on the line are output.

With reference to FIG. 4, the most general Viterbi algorithm will be described as a method of decoding a convolutionally encoded code.

On the decoding side, the signal corresponding to the bit line on the solid line or the broken line on the trellis diagram is received and the path on the trellis diagram is predicted to reproduce the original signal. However,
As will be described later, a delay corresponding to the path memory length occurs. As shown in the trellis diagram, there are two paths (branches) input to each state, and a 2-bit branch symbol based on the same rule as encoding is assigned to each branch.

First, when 2 bits are input, for each input branch to each state, a branch metric (metric) with the input bit is calculated, and the branch metric having a better value is selected. The sum of the branch metric accumulation (path metric) in the state before the selected branch is connected and the metric of the selected branch is taken as a new path metric in each state. Thus, each time the branch connected to each state is obtained, the path (path) information to reach each state is stored in the memory (path memory). Here, the cumulative result of selecting branches becomes a pass. Alternatively, the minimum unit of path is a branch.

When the above process is repeated for each 2-bit input,
According to the process of narrowing down the paths described in Document 3, the past paths are eventually narrowed down to one, so that a signal before convolutional coding is obtained from the obtained paths. Since the path memory length of the actual device is finite, if the paths do not converge even if the path memory length is exceeded, the path with the best path metric is selected at that point.

Next, the difference between the hard decision and the soft decision will be described.

A method of calculating a metric with a possible path on the trellis diagram by using the input bit value itself is called a hard decision. On the other hand, a method that uses the likelihood (likelihood) that an input bit value takes that value is called soft decision.
The soft decision has a higher accuracy of metric calculation and the bit error correction capability than the hard decision.

For example, not only in radio, but in digital signal transmission, in the case of hard decision, a certain reception level is set as a threshold value. To determine the signal value. On the other hand, in the case of soft decision, first, a seven-valued threshold value is set, the area is divided into eight areas according to the level of the received signal, and a value Ns of 0 to 7 is given to each area. That is, 1
An area that is sure to be 0, an area that is sure to be 0,
It is divided into a region that can take either 0 or 1, or a region closer to 1 rather. Here, the branch symbols 0 and 1 on the trellis diagram of FIG. 4 are set to −1 and 1, and the value Ns of 0 to 7 is converted to (2 × Ns−7) to calculate the sum of products of the input bit and the branch symbol. A Viterbi algorithm that selects branches with a large (correlation) becomes possible.

[0032]

However, the conventional decoding method has a problem in that the transmission bit string is fixed and the certainty of taking the phase is not taken into consideration.

That is, the modulated wave is affected by the frequency characteristic of the transmission path, the phase fluctuation with respect to the carrier wave, the frequency shift, and the like. The frequency characteristic of the transmission line is corrected by the equalizer 16, and the phase fluctuation and frequency shift with respect to the carrier wave are corrected by the phase synchronization circuit 17, respectively. However, when the distortion correction is successful depending on the degree of distortion of the modulated wave or the correction method. Then, there are cases where it does not work. If the correction is successful, the estimated bit string has a high probability of matching the actually transmitted bit string, but if the correction cannot be completed, the probability is small. It cannot be said that the credibility of the estimated bit string due to the success or failure of the correction depending on the transmission state is sufficiently reflected in the decoding.

The present invention has the problem that the above-mentioned conventional technique has a problem that the bit string is fixed at the time of decoding in the PSK modulation method and the possibility of actually transmitting the estimated bit string (bit likelihood) is not taken into consideration. A soft decision decoding method that solves the above problem is provided.

[0035]

In order to solve the above problems, the present invention calculates a bit likelihood from the deviation of the detected phase from the selected phase in the phase space and the distance from the origin, and softens this likelihood. It is characterized by being used for decision decoding.

That is, the phase / distance calculation processing for calculating the phase and the distance from the origin and the phase / distance calculation processing are obtained from the coordinate information on the phase space obtained by demodulating the signal modulated by the phase modulation method. The phase is compared with a unique phase transmitted by the phase modulation method, a phase selection process of selecting the unique phase having a small absolute value of the phase shift, and monotonically decreased with respect to the magnitude of the phase shift, And the absolute value of the phase shift is 0
, The maximum value of the likelihood is taken, and the absolute value of the phase shift is (2π
/ The number of unique phases transmitted by the phase modulation method), the phase likelihood calculation process for calculating the phase likelihoods is executed by the function taking the minimum value of the likelihoods.

Furthermore, as the distance from the origin increases, it increases monotonically, and if the distance from the origin is less than or equal to the first set value, the minimum value of the likelihood is taken and the distance from the origin becomes When the value is equal to or larger than the second set value, the phase likelihood correction process is performed to multiply the phase likelihood by the coefficient that takes the maximum value to obtain the corrected phase likelihood. Then, the corrected phase likelihood is used in the decoding process as a value indicating the certainty of taking a bit value.

[0038]

According to the present invention, since the soft decision decoding method is configured as described above, when the coordinate information on the phase space obtained by demodulating the signal modulated by the phase modulation method is input to the phase / distance calculation processing. In the phase / distance calculation process, the phase and distance are calculated from the input coordinate information, and the calculation result is sent to the phase selection process. In the phase selection processing, the input phase is compared with the unique phase transmitted by the phase modulation method, and the unique phase having a small absolute value of the phase shift is selected. In the phase likelihood calculation process, the phase likelihood is calculated based on the phase shift, and the calculation result is sent to the phase likelihood correction process. In the phase likelihood correction process, a corrected phase likelihood corresponding to the phase likelihood is calculated based on the distance. If this corrected phase likelihood is used for calculation of a metric or the like in soft decision decoding, an accurate reproduced signal can be obtained. Therefore, the above problem can be solved.

[0039]

1 is a flow chart of processing steps of a soft decision decoding method according to an embodiment of the present invention, and FIG. 5 is a configuration block diagram showing a receiving unit of a radio signal transmitting / receiving apparatus for implementing the soft decision decoding method. Is. In FIG. 5, elements common to those in the conventional FIG. 2 are designated by common reference numerals. The π / 4 shift DQPSK system is assumed as the modulation system.

First, the receiving section of the radio signal transmitting / receiving apparatus shown in FIG. 5 will be described.

This receiving section is provided with the same demodulating section 10 as that of the conventional one shown in FIG. 2, and the soft decision decoding section 30 is connected to the output side thereof. The soft-decision decoding unit 30 receives the coordinate information b on the phase space output from the demodulation unit 10 and corrects the bit error of the baseband signal to obtain the reproduction signal c.
It is composed of an individual circuit using a large scale integrated circuit (LSI) or the like, or program control using a processor.

Although the illustration of the transmission unit provided in the radio signal transmission / reception apparatus is omitted, the transmission unit encodes the original signal (block coding, convolutional coding, etc.) and performs interleave conversion. And π / 4 shift DQ
Modulate by PSK method. The present embodiment is effective even when the encoding or the interleave conversion is not performed, but the case where the encoding or the interleave conversion is performed is more general, and therefore is included in the drawing.

The soft-decision decoding unit 30 has a soft-decision data operation unit 40 which calculates soft-decision data from the coordinate information b. The soft decision decoding unit 40 includes a phase calculation unit 41, a phase selection unit 42, a phase likelihood calculation unit 43, a distance calculation unit 44, a distance coefficient calculation unit 45, a phase likelihood correction unit 46, a bit likelihood calculation unit 47, And a decision data output unit 48, and a deinterleave unit 50 is connected to the output side thereof.

The deinterleave unit 50 has a function of storing the soft decision data calculated by the soft decision data operation unit 40, and the decoding unit 60 is connected to the output side thereof. The data stored in the deinterleave unit 50 is read out by the decoding unit 60 while returning to the original bit order rearranged at the time of transmission. Therefore, the decoding unit 60 uses the read soft decision data. It has a function of performing a decoding process and outputting a reproduction signal c.

Next, the soft-decision decoding method of this embodiment will be described with reference to FIG.
Will be described with reference to.

In the demodulator 10 of FIG. 5, the RF band received signal a
Is input, the coordinates (that is, the in-phase component zp and the quadrature component zq) of the RF band received signal a in the phase space are detected, and the detected coordinate information b is stored in the soft decision decoding unit 30 in step S1. To the soft decision data calculation unit 40. The soft decision data calculation unit 40 performs the soft decision data processing S10 according to steps S11 to S18.

That is, when the coordinate information b is input, the phase calculation unit 41 calculates the phase in step S11. The phase zirad is given by equation (4). The calculated phase takes an arbitrary value of 0 to 2π. zirad = tan ⁻¹ (zq / zp) (4) In step S12, the phase selection unit 42 selects a phase. Since the π / 4 shift DQPSK system is assumed, if the detected phase signal is input at an odd number, one of the four phases (0, π / 2, π, 3π / 2) is selected. However, if the even number is input, (π / 4, 3
One of four phases (π / 4, 5π / 4, 7π / 4) is selected. t-th input phase selection value irad
(T) is the detected phase z as shown in the following equation (5).
Absolute value θ of the difference between irad and the candidate phase kπ / 4
Let (k) be the minimum phase. irad (t) = kπ / 4 | θ (k) is the minimum k (5) where θ (k) = | zirad−kπ / 4 | k = 0, 2, 4, 6 t = 2n + 1 = 1 , 3, 5, 7 t = 2n For k that satisfies the equation (5), θ (k) is 0 ≦ θ
(K) ≦ (number of unique phases transmitted by π / PSK method)
Meet The number of unique phases transmitted in the π / 4 shift DQPSK system is four.

In step S13, the phase likelihood calculator 43
Thus, the likelihood (likelihood) of taking the phase selected in step S12 is calculated. Likelihood is the selected phase kπ / 4
And a phase shift θ (k) of the detected phase zirad,
It is expressed as the following equation (6). Selected phase irad
The likelihood of taking (t) is prad (t). prad (t) = (cos2θ (k) +1) / 2 (6) In step S14, the distance calculation unit 44 calculates the distance zz from the origin from the detected coordinates (zp, zq) as shown in equation (7). calculate. zz ² = zp ² + zq ² (7) Next, in step S15, the distance coefficient calculator 45 causes the distance coefficient plen (t) to be calculated according to the distance from the origin.
Is calculated from the equation (8). plen (t) = 0 (zz ² ≦ zz ² _th1 ) = 1 (zz ² > zz ² _th2 ) ... (8) However, in the present embodiment, the set values zz _th1 and zz _th2 are set to the same value, and by the AGC 13, Assuming that the average power is set to 1, zz _th1 and zz _th2 are defined as follows. _{zz th1 = zz th2 = 0.24 zz} 2 th1 = zz 2 th2 = 0.0576 detected coordinates (zp, zq), detected phase Zirad, the distance from the origin zz, phase irad selected (t), the set value zz The relationship of _th1 is shown in FIG.

FIG. 6 shows the case where t = 2n + 1 and the detected phase z
From the value of irad, irad (t) = 0. In addition,
When the detection phase zirad has the same value and t = 2n, i
rad (t) = π / 4. The expression (8) means that when the coordinates are inside the chain double-dashed line, this coordinate information is not trusted and only when it is outside the chain double-dashed line, it is treated as valid information. The solid line circle indicates the average power by the AGC 13.

In step S16, the phase likelihood correction unit 46
From the distance coefficient obtained in step S15,
As in Expression (9), the likelihood prad (t) in Expression (6) is corrected to obtain prad2 (t). prad2 (t) = prad (t) * plen (t) (9) In the case of the π / 4 shift DQPSK system, the phase difference between the phase selected in step S16 and the phase selected immediately before is (10 ) It calculates like a formula. The previously selected phase and its likelihood are stored in the memory in the phase likelihood correction unit 46. idif = irad (t) -irad (t-1) (10) The likelihood of the phase difference is calculated as the lower likelihood of the phase likelihoods at consecutive time points as in the following Expression (11). .. However, in the present embodiment, the calculation of the phase difference likelihood is not limited to the equation (11). pdif = min (prad (t), prad (t-1)) (11) In step S17, the bit likelihood operation unit 47 determines the bit string corresponding to the phase difference selected as described above to the bit likelihood. It is read from the memory in the arithmetic unit 47. Phase difference π
/ 4 (0, 0), 3π / 4 (0, 1), 5π /
The case of 4 (1, 1) corresponds to the case of 7π / 4 (1, 0). The bits corresponding to the phase difference idif are sequentially ib1, i
b2. ib1 and ib2 take 0 or 1.

The bit likelihoods pb1 and pb2 are calculated by the following equation (12). This bit likelihood is considered to be -1 to 1. pb1 = 1-2 pdif ib1 = 0 = 2 pdif-1 ib1 = 1 pb2 = 1-2 pdif ib2 = 0 = 2 pdif-1 ib2 = 1 pb1 and pb2 obtained by the above process The likelihood indicates that the bit value may be 0 or 1. p
b1 takes an arbitrary value between -1 and 1, and is highly likely to be 1 when pb1 is close to 1, and is likely to be 0 when pb1 is close to -1. The same applies to pb2.

In the final step S18 of the soft decision data processing S10, the soft decision data is output from the decision data output unit 48 and sent to the decoding unit 60 via the deinterleave unit 50. In step S20, the decoding unit 60 performs decoding processing using the input soft decision data and outputs the reproduction signal c. As the decoding process, for example, when the code is convolutional coded, Viterbi decoding is performed. High-precision decoding is possible even for block coding.

FIG. 7 shows a simulation result of bit error characteristics when the soft decision decoding method of this embodiment is applied to Viterbi decoding. The horizontal axis represents the ratio Eb / No of the average signal energy Eb per one bit to the noise power density No, and the vertical axis represents the bit error rate. The curves in the figure are those in which Δ is not used for convolutional coding, □ is for soft-decision Viterbi decoding of a conventional convolutional code, and ◯ is for soft-decision Viterbi decoding according to the present embodiment.

The simulation conditions of FIG. 7 will be described. 171 bit original signal per slot is class 1
(89 bits) and class 2 (82 bits) are divided, and only the class 1 signal is convolutionally encoded. Coding rate 1/2 for convolutional coding, constraint length 6, generator polynomial 11
0101 and 101111. After tamper coding,
Class 1 signal (178 bits) and Class 2 signal (8
(2 bits) is interleaved by a 26 × 10 array and modulated by the π / 4 shift DQPSK method, and then an error is randomly given to the phase information. On the receiving side, the phase information is converted into bit information (bit likelihood in the case of soft decision), deinterleaved conversion is performed, and only class 1 is Viterbi-decoded. The above processing is executed for 200 slots, and class 1 and class 2
For each of, calculate the bit error rate.

As is apparent from FIG. 7, in the present embodiment, when transmitting with the same Eb / No, the bit error rate becomes smaller than that in the conventional hard decision Viterbi decoding. Conversely, when it is desired to have the same bit error rate, less transmission power is required.

The present embodiment is not limited to the above embodiment, but various modifications can be made. Examples of such modifications include the following.

(1) In step S12 of FIG. 1, the calculation formula of the phase likelihood is not limited to the formula (6), but it decreases monotonically with respect to the magnitude of θ (k), and θ (k) is 0. In the case of, the maximum value of the likelihood is taken, and when θ (k) is (2π / the number of unique phases transmitted in the differential PSK method), a function taking the minimum value of the likelihood may be used. In the above embodiment, the likelihood is considered to be 0 to 1, but it may be 0 to 100, or -1 to 1, depending on the case.

(2) In step S18 of FIG. 1, the bit likelihood obtained in step S17 may be used as it is as the soft decision data. However, since the bit likelihood in step S17 is a real number, appropriate quantization is performed to obtain this bit likelihood. The value may be soft decision data. In addition, this bit likelihood can be used as it is as the metric calculation method of the Viterbi algorithm when the product-sum calculation is performed, and can be diverted with a slight change when calculating the metric by another method such as difference calculation. Is.

(3) Soft decision data calculation processing S10 in FIG.
Can also be used for a soft decision decoding method of a block code. Also in the case of soft-decision decoding of a block code, the bit error correction capability is improved in the calculation by the bit likelihood for the same reason as in the case of soft-decision Viterbi decoding. Further, it is possible to use both block code decoding and Viterbi decoding together.

(4) The Viterbi decoding shown in FIG. 1 can be used in combination with various diversity receptions. Combination with decision feedback type equalization is also possible. In addition to block code and interleave, ARQ (AUTOMATIC REPEAT)
REQUEST type code error control (a method of retransmitting information when an error is detected) can be used together.

(5) The calculation formula of the amplitude coefficient plen (t) in step S15 of FIG. 1 is not limited to the formula (8), and increases monotonically with the magnitude of the distance zz, and zz is set to the first setting. A function that takes the minimum value of the likelihood when the value is less than or equal to the value and takes the maximum value of the likelihood when zz is equal to or greater than the second set value may be used.

[0062]

As described above in detail, according to the present invention, in the phase modulation method, in the process of obtaining the bit likelihood from the received carrier, the deviation of the detected phase from the selected phase in the phase space and Bit likelihood is calculated from the distance from the origin, and this likelihood is used for soft-decision decoding, so the bit error rate of the original signal can be reduced compared to conventional hard-decision decoding, and highly accurate decoding is possible. Can be done.

[Brief description of drawings]

FIG. 1 is a flowchart showing processing steps of a soft decision decoding method according to an embodiment of the present invention.

FIG. 2 is a configuration block diagram of a receiving unit in a conventional wireless signal transmitting / receiving apparatus.

FIG. 3 is an explanatory diagram of convolutional coding.

FIG. 4 is a diagram showing a trellis figure.

FIG. 5 is a configuration block diagram of a receiving unit in the wireless signal transmitting / receiving apparatus showing the embodiment of the present invention.

FIG. 6 is a diagram showing a relationship between detected coordinates and selected phases in FIG.

FIG. 7 is a bit error characteristic diagram of FIG.

[Explanation of symbols]

 10 demodulation unit 30 soft decision decoding unit 40 soft decision data calculation unit 41 phase calculation unit 42 phase selection unit 43 phase likelihood calculation unit 44 distance calculation unit 45 distance coefficient calculation unit 46 phase likelihood correction unit 47 bit likelihood calculation unit 48 Judgment data output unit 50 Deinterleave unit 60 Decoding unit a RF band received signal b Coordinate information c Reproduced signal S1 Coordinate detection process S10 Soft decision data process S11 Phase calculation process S12 Phase selection process S13 Phase likelihood calculation process S14 Distance calculation process S15 Distance coefficient calculation process S16 Phase likelihood correction process S17 Bit likelihood calculation process S18 Decision data output process S20 Soft decision decoding implementation process

Claims (1)

Claim: What is claimed is: 1. A phase / distance calculation process for calculating a phase and a distance from an origin from coordinate information on a phase space obtained by demodulating a signal modulated by a phase modulation method, and the phase / distance calculation process. The phase obtained by the distance calculation process is compared with the unique phase transmitted by the phase modulation method, and the phase selection process of selecting the unique phase having a small absolute value of the phase deviation, and the magnitude of the phase deviation. On the other hand, when it decreases monotonically and the absolute value of the phase shift is 0, the maximum value of the likelihood is taken, and when the absolute value of the phase shift is (2π / the number of unique phases transmitted by the phase modulation method), A phase likelihood calculation process for calculating each phase likelihood by a function that takes the minimum value of the likelihood, and the distance from the origin increases monotonically as the distance from the origin increases, and the distance from the origin is set to the first setting. If it is less than or equal to the value, the minimum likelihood is taken and When the distance from the point is equal to or larger than the second set value, a phase likelihood correction process of obtaining a corrected phase likelihood by multiplying the phase likelihood with a coefficient that takes the maximum value is performed, and the corrected phase likelihood is A soft-decision decoding method characterized by being used in a decoding process as a value indicating the probability of taking a bit value.