JP7660970B2 - Log likelihood ratio calculation circuit and wireless receiving device - Google Patents

Description

This invention relates to a log-likelihood ratio calculation circuit and a wireless receiving device, and in particular to a circuit that calculates a log-likelihood ratio used in error correction in low-density parity check decoding and a wireless receiving device that includes the circuit.

Conventionally, in low density parity check (LDPC) decoding, a method for performing error correction using a log-likelihood ratio (LLR: Log-Likelihood Ratio) calculated based on the distribution of thermal noise in the received signal is known (see, for example, Patent Document 1).

JP 2013-201582 A

However, in high-multiple-level modulation methods, there is a problem in that the error correction capability of the low-density parity check is reduced due to the large influence of factors other than thermal noise, such as phase noise and fading of the received signal. Specifically, the conventional method of calculating the log-likelihood ratio assumes that the amplitude of the received signal fluctuates based on a normal distribution/Gaussian distribution centered on the ideal symbol point due to thermal noise in the received signal. However, this method cannot take into account the fluctuation that occurs when the phase of the symbol rotates due to phase noise and fading of the received signal. For this reason, the calculation accuracy of the log-likelihood ratio deteriorates, especially in an environment where phase noise is present, and ultimately the error correction capability of the low-density parity check is reduced.

The present invention aims to provide a log-likelihood ratio calculation circuit and a wireless receiving device including the log-likelihood ratio calculation circuit that can improve the accuracy of calculating the log-likelihood ratio and thereby improve the error correction capability of the low-density parity check.

In order to solve the above problem, the invention described in claim 1 is a log likelihood ratio calculation circuit that calculates a log likelihood ratio by multiplying a log likelihood ratio based on thermal noise and a log likelihood ratio based on phase noise by a weighting factor and then adding them up, and calculates the log likelihood ratio based on the phase noise based on a probability density distribution that becomes 0 and a probability density distribution that becomes 1 according to the phase angle of an ideal symbol point that becomes 0 or 1 according to the amplitude of the ideal symbol point ("amplitude discrimination signal") that becomes 0 or 1 according to the amplitude of the ... and an ideal symbol point (called a "phase-distinguishing symbol point") which is distinguished as being 0 or 1 depending on a phase angle, the circuit determines whether the bit is the amplitude-distinguishing symbol point or the phase-distinguishing symbol point based on the amplitude of the received signal, and then calculates the amplitude-distinguishing symbol point based on a probability density distribution which becomes 0 and a probability density distribution which becomes 1 depending on the amplitude using the amplitude of the received signal, and calculates the phase-distinguishing symbol point based on a probability density distribution which becomes 0 and a probability density distribution which becomes 1 depending on the phase angle using the phase angle of the received signal .

According to a second aspect of the present invention, in the log-likelihood ratio calculation circuit according to the first aspect, a log-likelihood ratio based on the phase noise of each bit is calculated according to a predetermined formula.

A third aspect of the present invention provides a radio receiving device comprising the log-likelihood ratio calculation circuit according to the first or second aspect of the present invention.

According to the invention described in claims 1 and 2 , the log-likelihood ratio based on thermal noise and the log-likelihood ratio based on phase noise are taken into consideration of the degree of mutual priority between them (specifically, after multiplying each log-likelihood ratio by a weighting coefficient), and then these are added to calculate the log-likelihood ratio for each bit.This takes into account not only the influence of thermal noise, but also factors such as phase noise and distribution fluctuations due to fading when calculating the log-likelihood ratio, making it possible to realize a low-density parity check that is robust against disturbances including phase noise.

According to the inventions described in claims 1 and 2 , the log-likelihood ratio is calculated taking into account the probability distribution when the phase of the symbol rotates or fluctuates based on a normal distribution/Gaussian distribution due to phase noise or fading of the received signal, etc., so that it is possible to improve the calculation accuracy of the log-likelihood ratio even in an environment where phase noise of the received signal is dominant, and ultimately to improve the error correction capability of the low-density parity check.

According to the third aspect of the present invention, the above-mentioned advantageous effects can be achieved in a wireless receiving device that performs error correction in low-density parity check decoding using a log-likelihood ratio.

1 is a functional block diagram showing a schematic configuration of a wireless receiving device according to an embodiment of the present invention, including an LLR calculation unit as a log-likelihood ratio calculation circuit. 1 is a functional block diagram showing a schematic configuration of an LLR calculation unit serving as a log-likelihood ratio calculation circuit according to an embodiment; 1 is a diagram showing an arrangement of ideal symbol points in a 16QAM system and bit sequences assigned to each ideal symbol point; FIG. 13 is a diagram for explaining the setting of variables related to the calculation of a log-likelihood ratio. FIG. 13 is a diagram for explaining the concept of a calculation formula for a log-likelihood ratio of the 1st bit, and in particular, a diagram for explaining a region where the 1st bit is 0 and a region where the 1st bit is 1. FIG. 13 is a diagram for explaining the concept of a calculation formula for a log-likelihood ratio of the first bit. FIG. 13 is a diagram for explaining the concept of a calculation formula for the log-likelihood ratio of the second bit, and in particular, a diagram for explaining a symbol having bit information carried on the amplitude and a symbol having bit information carried on the phase. FIG. 13 is a diagram for explaining the concept of a calculation formula for the log-likelihood ratio of the second bit. 1A is a diagram for explaining the concept of a weighting coefficient of a log-likelihood ratio, and FIG. 1B is a diagram for explaining a case where a log-likelihood ratio based on thermal noise is prioritized, and FIG. 1B is a diagram for explaining a case where a log-likelihood ratio based on phase noise is prioritized. 1 is a diagram for explaining problems with a conventional method for calculating a log-likelihood ratio and a countermeasure by a log-likelihood ratio calculation circuit according to the present invention; FIG. 1 is a functional block diagram showing a schematic configuration of an evaluation system used in a verification example of the effectiveness of a log-likelihood ratio calculation circuit according to the present invention. 13 is a diagram showing a code error rate as an evaluation result in a verification example of the effectiveness of the log-likelihood ratio calculation circuit according to the present invention. FIG.

The present invention will be described below based on the illustrated embodiment. Note that the following describes the characteristic configuration of the present invention, and the description of the conventional mechanism for wireless communication will be omitted. In addition, in each figure, the signal line transmitting the real part (I signal; in other words, the in-phase component, I signal component) and the signal line transmitting the imaginary part (Q signal; in other words, the quadrature component, Q signal component) of the complex signal are shown together as a single signal line. In addition, in the following description, "greater than or equal to" and "greater than" may be interchangeable, and furthermore, "less than or equal to" and "less than" may be interchangeable.

Figure 1 is a functional block diagram showing a schematic configuration of a wireless receiving device 1 according to an embodiment of the present invention, including an LLR calculation unit 20 as a log-likelihood ratio calculation circuit.

The antenna 10 receives a signal wave sent from a wireless transmitting device (not shown) and transfers the signal wave to the channel filter 11 as a received signal.

The wireless transmitting device performs low-density parity check coding on the wireless frame to be transmitted and modulates it using a quadrature amplitude modulation method with a multi-level of 16 (i.e., 16QAM method; QAM is an abbreviation for Quadrature Amplitude Modulation), and then transmits it from the antenna. In other words, the wireless receiving device 1 receives a multi-level modulated signal that has been subjected to low-density parity check coding and modulated using a quadrature amplitude modulation method with a multi-level of 16 (16QAM method).

The channel filter 11 is a filter for limiting the reception band, and specifically may be configured, for example, by a bandpass filter. The channel filter 11 receives the reception signal transferred from the antenna 10, performs band limiting processing on the reception signal, and extracts and outputs the reception signal of the desired frequency band from the reception signal.

The mixer 13 receives the received signal output from the channel filter 11 and the local signal output from the local oscillator 12, multiplies the received signal by the local signal, and converts (specifically, down-converts) the frequency of the received signal and outputs it.

The automatic gain control unit 14 receives the received signal output from the mixer 13, performs gain adjustment processing on the received signal so that the level of the received signal is constant at a predetermined level, and outputs the received signal (which is an analog signal) after gain adjustment processing (in other words, after level correction).

The A/D converter 15 (Analog-to-Digital converter) receives the received signal output from the automatic gain control unit 14, and performs analog-to-digital conversion processing on the received signal to convert it from an analog signal to a digital signal. In other words, the A/D converter 15 outputs a digital received signal.

The digital quadrature detection unit 16 receives the digital received signal output from the A/D converter 15 and converts the received signal into a baseband signal by quadrature detection. Specifically, the digital quadrature detection unit 16 includes a multiplier that receives a cosine wave from a numerically controlled oscillator and synchronously detects the in-phase component of the received signal, and a multiplier that receives a sine wave from the numerically controlled oscillator and synchronously detects the quadrature component of the received signal. The digital quadrature detection unit 16 performs quadrature detection of multi-level modulation by digital signal processing, and generates and outputs a complex signal (in other words, a detection signal of the in-phase component and the quadrature component) expressed in IQ quadrature coordinates with the real component (I signal) constituting the real part and the imaginary component (Q signal) constituting the imaginary part.

The numerically controlled oscillator of the digital quadrature detection unit 16 is an oscillator that oscillates at a frequency according to a set value, and generates a cosine wave and a sine wave whose phases differ by 90° from each other, and supplies them to each of the two multipliers.

The roll-off filter 17 receives the baseband signal (in other words, the detection signal of the in-phase component and the quadrature component) output from the digital quadrature detection unit 16, processes the in-phase component and the quadrature component of the baseband signal to remove inter-symbol interference, and shapes the spectrum waveform of the baseband signal to have the desired roll-off before outputting it.

The equalizer 18 receives the baseband signal output from the roll-off filter 17, performs adaptive equalization processing on each of the in-phase and quadrature components of the baseband signal to remove inter-symbol interference caused by frequency selective fading, and outputs the baseband signal after adaptive equalization processing (received signal).

The decoding unit 19 receives the log-likelihood ratio (LLR: abbreviation for Log-Likelihood Ratio) output from the LLR calculation unit 20, and performs low-density parity check decoding processing using the log-likelihood ratio, for example, according to the sum-product decoding method.

The LLR calculation unit 20 is a circuit that calculates the log-likelihood ratio used in error correction in low-density parity check decoding, and is incorporated into the wireless receiving device 1.

Figure 2 is a functional block diagram showing the schematic configuration of the LLR calculation unit 20 as a log-likelihood ratio calculation circuit according to the embodiment.

The LLR calculation unit 20 according to the embodiment multiplies the log-likelihood ratio LLR _{N(bit_num)} based on thermal noise and the log-likelihood ratio LLRφ _{(bit_num)} based on phase noise by weighting coefficients W _N and Wφ, respectively, and then adds them together to calculate the log-likelihood ratio LLR _{bit_num} .

The LLR calculation unit 20 according to the embodiment further calculates a log likelihood ratio LLRφ _{(bit_num)} based on phase noise, for bits that are distinguished as being 0 or 1 depending on the phase angle φ _i of an ideal symbol point, of a 4-bit received signal modulated by a quadrature phase amplitude modulation method with a multi-level number of 16 on the transmitting side, based on a probability density distribution that becomes 0 and a probability density distribution that becomes 1 depending on the phase angle, using the phase angle xφ of the received signal, and for bits that are classified into an ideal symbol point that is distinguished as being 0 or 1 depending on the amplitude (referred to as an "amplitude-distinguishing symbol point") and an ideal symbol point that is distinguished as being 0 or 1 depending on the phase angle (referred to as a "phase-distinguishing symbol point"), it determines whether the bit is an amplitude-distinguishing symbol point or a phase-distinguishing symbol point based on the amplitude x _R of the received signal, and then calculates the log likelihood ratio LLRφ(bit_num) based on the phase noise, using the phase angle xφ of the received signal. Using _R , calculations are made based on a probability density distribution which becomes 0 and a probability density distribution which becomes 1 depending on the amplitude, and for the phase-distinguishing symbol points, using the phase angle xφ of the received signal, calculations are made based on a probability density distribution which becomes 0 and a probability density distribution which becomes 1 depending on the phase angle.

First, the arrangement of ideal symbol points in the 16QAM system and the bit strings assigned to each ideal symbol point (in other words, the signal space diagram of 16QAM) are shown in Fig. 3. In Fig. 3, A _i is the in-phase component of the ideal symbol point arrangement, and B _i is the quadrature component of the ideal symbol point arrangement (where i = 1, 2, 3, 4).

The LLR calculation unit 20 calculates the log-likelihood ratio LLR _{bit_num} of the (bit_num)th bit according to the following equation 1.
(Math. 1) LLR _{bit_num} = W _N ×LLR _{N(bit_num)} +Wφ×LLRφ _{(bit_num)}
Here, bit_num: bit number (where bit_num = 1, 2, 3, 4)
LLR _{N(bit_num)} : Log-likelihood ratio based on thermal noise
LLRφ _{(bit_num)} : Log-likelihood ratio based on phase noise
W _N : Weighting factor of log-likelihood ratio based on thermal noise
Wφ: weighting factor of log-likelihood ratio based on phase noise

Specifically, the log-likelihood ratio LLR _{N(bit_num)} based on thermal noise of the (bit_num)th bit is calculated according to, for example, the following Equations 2A to 2D.

The meanings of the symbols and variables in the above formulas 2A to 2D are as follows.
LLR _N1 : Log likelihood ratio based on thermal noise of the 1st bit LLR _N2 : Log likelihood ratio based on thermal noise of the 2nd bit LLR _N3 : Log likelihood ratio based on thermal noise of the 3rd bit LLR _N4 : Log likelihood ratio based on thermal noise of the 4th bit x _i : In-phase component of the received signal x _q : Quadrature component of the received signal A _i : In-phase component of the ideal symbol point constellation (where i = 1, 2, 3, 4)
B _i : Orthogonal components of the ideal symbol point constellation (where i=1, 2, 3, 4)
σ ² : Noise variance

Specifically, the log-likelihood ratio LLRφ _{(bit_num)} based on the phase noise of the (bit_num)th bit is calculated, for example, according to the following formulas 3 to 6. Here, the settings of variables used in the following formulas 3 to 6, which are defined for the arrangement of ideal symbol points shown in FIG. 3, are shown in FIG.

The meanings of the symbols and variables in the following Equations 3 to 7 are as follows: Note that the phase angle "φ" of the ideal symbol point in the following equations is represented as "Φ" in the drawings.
xφ: Phase angle of the received signal x _R : Amplitude of the received signal φ _i : Phase angle of the ideal symbol point (where the subscript i of φ is 0, 1, 2, ..., 11)
R _i : Amplitude of the ideal symbol point (where the subscript i of R = 0, 1, 2)
σ ² : Noise variance R _thre1 : First amplitude threshold R _thre2 : Second amplitude threshold

The log-likelihood ratio LLRφ ₁ based on the phase noise of the first bit is calculated according to the following Equation 3.

The log-likelihood ratio LLRφ2 based on the phase noise of the second bit is calculated according to the following formula _4A when the amplitude _xR of the received signal is _equal to or greater than the first amplitude threshold _Rthre1 or equal to or less than the second amplitude threshold Rthre2 (i.e., _xR ≧ _Rthre1 or _xR ≦ _Rthre2 ), and is calculated according to the following formula 4B when the amplitude _xR of the received signal is less than the first amplitude threshold _Rthre1 and greater than the second amplitude threshold _Rthre2 (i.e., _Rthre2 < _xR < _Rthre1 ).

The log-likelihood ratio LLRφ ₃ based on the phase noise of the third bit is calculated according to the following Equation 5.

The log-likelihood ratio LLRφ4 based on the phase noise of the 4th bit is calculated according to the following formula _6A when the amplitude _xR of the received signal is _equal to or greater than the first amplitude threshold _Rthre1 or equal to or less than the second amplitude threshold Rthre2 (i.e., _xR ≧ _Rthre1 or _xR ≦ _Rthre2 ), and is calculated according to the following formula 6B when the amplitude _xR of the received signal is less than the first amplitude threshold _Rthre1 and greater than the second amplitude threshold _Rthre2 (i.e., _Rthre2 < _xR < _Rthre1 ).

The above formulas 3 to 6 are based on the following idea (see Figures 5 to 8).

The first bit is ₀ when the phase angle is -90° to +90° (in other words, 270° to 90° through 360°/0°; that is, the fourth and first quadrants of the IQ orthogonal coordinate system including the phase angles of the ideal symbol points Φ ₉ to Φ ₁₁ , Φ 0 to Φ ₂ in FIG. 5 ), and is 1 when the phase angle is +90° to -90° (in other words, 270° through 180°; that is, the second and third quadrants of the IQ orthogonal coordinate system including the phase angles of the ideal symbol points Φ ₃ to Φ ₅ , Φ ₆ to Φ ₈ in FIG. 5 ). Therefore, if it is considered that the phase of the received signal fluctuates based on a normal distribution/Gaussian distribution, the probability density of the first bit being 0 is shown as in FIG. 6A, and the probability density of the first bit being 1 is shown as in FIG. 6B. Based on this, by considering these probability density distributions as each term, the above-mentioned formula 3 for calculating the log-likelihood ratio LLRφ ₁ based on the phase noise of the 1st bit is derived (see FIG. 6C).

The third bit is based on the same idea as the first bit, that is, the third bit becomes 0 when the phase angle goes from 0° through +90° to +180° (i.e., the first and second quadrants of the IQ orthogonal coordinate system including the phase angles of the ideal symbol points Φ ₀ through Φ ₂ , Φ ₃ to Φ ₅ in FIG. 5 ), and becomes 1 when the phase angle goes from +180° through +270° to 0° (i.e., the third and fourth quadrants of the IQ orthogonal coordinate system including the phase angles of the ideal symbol points Φ ₆ through Φ ₈ , Φ ₉ to Φ ₁₁ in FIG. 5 ). Based on this, the above-mentioned Equation 5 for calculating the log-likelihood ratio LLRφ ₃ based on the phase noise of the third bit is derived by considering the probability density distribution in which the third bit is 0 and the probability density distribution in which the third bit is 1 as each term.

As for the second bit, the same concept can be applied to all quadrants of the IQ orthogonal coordinate system, so the first quadrant of the IQ orthogonal coordinate system will be taken as an example for explanation. As shown in FIG. 7, the second bit is divided into an ideal symbol point where the amplitude determines whether it is 0 or 1 (the "symbol with bit information on the amplitude" in the figure; this is an amplitude-distinguishing symbol point), and an ideal symbol point where the phase determines whether it is 0 or 1 (the "symbol with bit information on the phase" in the figure; this is a phase-distinguishing symbol point). For this reason, the amplitude _xR of the received signal is compared with the first amplitude threshold _Rthre1 and the second amplitude threshold _Rthre2 to determine whether the received signal/symbol has bit information on the amplitude or phase, and the method of calculating the log-likelihood ratio is changed for each. For a symbol carrying bit information in the amplitude, the probability density distribution in which the second bit is 0 and the probability density distribution in which the second bit is 1 are shown in Fig. 8A, and for a symbol carrying bit information in the phase, the probability density distribution in which the second bit is 0 and the probability density distribution in which the second bit is 1 are shown in Fig. 8B. Based on this, by considering these probability density distributions as each term, the above-mentioned formula 4 for calculating the log-likelihood ratio _LLRφ2 based on the phase noise of the second bit is derived (see Fig. 8C).

The fourth bit is based on the same idea as the second bit, and in the same way as the second bit, for symbols which carry bit information in amplitude, and for symbols which carry bit information in phase, the fourth bit is 0 at phase angles of the ideal symbol point _Φ2 , _Φ3 , _Φ8 , and _Φ9 and 1 at phase angles of the ideal symbol point _Φ0 , _Φ5 , _Φ6 , and _Φ11. By considering the probability density distribution where the fourth bit is 0 and the probability density distribution where the fourth bit is 1 for symbols which carry bit information in amplitude, and the probability density distribution where the fourth bit is 0 and the probability density distribution where the fourth bit is 1 for symbols which carry bit information in phase, as each term, the above formula 6 for calculating the log-likelihood ratio _LLRφ4 based on the phase noise of the fourth bit can be derived.

The first amplitude threshold _Rthre1 and the second amplitude threshold _Rthre2 are not limited to specific values, and for example, the first amplitude threshold _Rthre1 is set between the amplitudes _R1 and _R2 of the ideal symbol points, and the second amplitude threshold _Rthre2 is set between the amplitudes _R0 and _R1 of the ideal symbol points, and are appropriately set to appropriate values while taking into consideration that a symbol having bit information in its amplitude and a symbol having bit information in its phase can be well distinguished. Specifically, for example, the first amplitude threshold _Rthre1 may be set as shown in the following formula 7A, and the second amplitude threshold _Rthre2 may be set as shown in the following formula 7B.
(Math. 7A) R _thre1 = R ₁ + (R ₂ - _{R 1} )/2
(Math. 7B) R _thre2 = R _o +(R ₁ - _{R 0} )/2

Then, by performing low-density parity check decoding processing using the log-likelihood ratios calculated according to the above-mentioned formula 1, the log-likelihood ratios LLRφ based on the distribution of phase noise are added to the log-likelihood ratios LLR _N based on the distribution of thermal noise, thereby improving robustness against phase noise. Also, since the influence of phase noise differs depending on the magnitude of thermal noise and the position of the symbol, the weighting coefficients W _N and Wφ of each log-likelihood ratio control whether the log-likelihood ratio based on thermal noise or phase noise is to be prioritized.

The weighting coefficients _WN , Wφ of each log-likelihood ratio are not limited to specific values, but are appropriately set to appropriate values after taking into consideration the degree to which the log-likelihood ratio based on thermal noise or phase noise is to be prioritized, taking into account that the influence of phase noise varies depending on the magnitude of thermal noise and the position of the symbol.

The weighting coefficients W _N , Wφ of each log-likelihood ratio may be calculated by a predetermined calculation formula. Specifically, for example, the weighting coefficient W _N of the log-likelihood ratio based on thermal noise may be calculated according to the following formula 8A, and the weighting coefficient W _N of the log-likelihood ratio based on phase noise may be calculated according to the following formula 8B.

The meanings of the symbols and variables in the above formulas 8A and 8B are as follows:
N: Variance of the thermal noise (i.e., √N corresponds to the average amplitude of the thermal noise)
S _d : Symbol interval in QAM system α: Priority adjustment coefficient β: Equivalent adjustment coefficient

The priority adjustment coefficient α is a coefficient for determining the percentage or less of the symbol interval Sd of the QAM system at which the log-likelihood ratio LLRφ based on phase noise is prioritized when the average value of the amplitude fluctuation due to thermal noise (i.e., √N ₎ is less than or equal to that percentage.

The priority adjustment coefficient α is not limited to a specific value, but is set to an appropriate value after taking into consideration the degree of priority between thermal noise and phase noise. Specifically, the priority adjustment coefficient α may be set to a value in the range of about 0.1 to 0.5, for example, as one example only.

The equality adjustment coefficient β is a coefficient for equalizing the magnitude of each log-likelihood ratio when the log-likelihood ratio due to thermal noise and the log-likelihood ratio due to phase noise have equal priority.

The equality adjustment coefficient β is not limited to a specific value, but is appropriately set to a value that equalizes the magnitude of each log-likelihood ratio when the log-likelihood ratio due to thermal noise and the log-likelihood ratio due to phase noise have equal priority, assuming the value of the priority adjustment coefficient α.

The priority adjustment coefficient α and the equal adjustment coefficient β are set in advance in the weighting coefficient control unit 24 of the LLR calculation unit 20.

By calculating the weighting coefficients _WN and Wφ of each log-likelihood ratio according to the above formula 8A and formula 8B, as shown in Fig. 9A, when the average amplitude √N of thermal noise is larger than α×100[%] of the symbol interval Sd of the _QAM method, the weighting coefficient _WN for the log-likelihood ratio based on thermal noise becomes large (specifically, becomes larger than 1), and conversely, the weighting coefficient Wφ for the log-likelihood ratio based on phase noise becomes small (specifically, becomes smaller than 1), that is, the log-likelihood ratio based on thermal noise is prioritized to calculate the final log-likelihood ratio LLR _{bit_num} . Note that in Fig. 9, α is set to 0.5 as an example only.

On the other hand, as shown in FIG. 9B , when the average amplitude √N of the thermal noise is equal to or less than α×100 [%] of the inter-symbol interval _Sd of the QAM system, the weighting factor Wφ for the log-likelihood ratio based on the phase noise becomes large (specifically, becomes 1 or more), and conversely, the weighting factor _WN for the log-likelihood ratio based on the thermal noise becomes small (specifically, becomes 1 or less), that is, the log-likelihood ratio based on the phase noise is prioritized in calculating the final log-likelihood ratio LLR _{bit_num} .

Note that, in Equation 8, the weighting factor _WN for the log-likelihood ratio based on thermal noise and the weighting factor Wφ for the log-likelihood ratio based on phase noise are calculated based on the magnitude of the thermal noise (in other words, with the thermal noise as a variable), but the weighting factor _WN for the log-likelihood ratio based on thermal noise and the weighting factor Wφ for the log-likelihood ratio based on phase noise may be calculated based on the magnitude of the phase noise (in other words, with the phase noise as a variable).

The thermal noise LLR calculation unit 21 of the LLR calculation unit 20 receives an input of a baseband signal (received signal) output from the equalizer 18, acquires an in-phase component x _i and a quadrature component x _q of the received signal from the baseband signal, and calculates and outputs a log-likelihood ratio LLR _{N(bit_num)} based on the thermal noise of the (bit_num)th bit in accordance with the above-mentioned Formula 2A to Formula 2D.

The phase noise LLR calculation unit 22 receives an input of a baseband signal (received signal) output from the equalizer 18, acquires a phase angle xφ and an amplitude _xR of the received signal from the baseband signal, and calculates and outputs a log-likelihood ratio LLRφ( _{bit_num} ) based on the phase noise of the (bit_num)th bit according to the above-mentioned Formulas 3 to ₆ , while taking into consideration the magnitude relationship between the amplitude _xR of the received signal and the first amplitude threshold R thre1 and the second amplitude threshold R _thre2 .

The thermal noise power estimation unit 23 receives the baseband signal (received signal) output from the equalizer 18, estimates the thermal noise power in the wireless receiving device 1 based on the baseband signal, and outputs an estimated amount of thermal noise power.

The weighting coefficient control unit 24 receives an input of the baseband signal (received signal) output from the equalizer 18, and also receives an input of the estimated amount of thermal noise power output from the thermal noise power estimating unit 23, and uses the baseband signal and the estimated amount of thermal noise power (specifically, calculates the variance N of the thermal noise) to calculate a weighting coefficient _WN of the log-likelihood ratio based on thermal noise according to the above Equation 8A, and also calculates and outputs a weighting coefficient Wφ of the log-likelihood ratio based on phase noise according to the above Equation 8B.

The first multiplier 25 receives an input of the log-likelihood ratio LLR _{N(bit_num)} based on thermal noise output from the thermal noise LLR calculation unit 21, and also receives an input of the weighting coefficient W _N of the log-likelihood ratio based on thermal noise output from the weighting coefficient control unit 24, multiplies the log-likelihood ratio LLR _{N(bit_num)} based on thermal noise by the weighting coefficient W _N of the log-likelihood ratio based on thermal noise, and outputs the multiplication result (i.e., W _N × LLR _{N(bit_num)} ).

The second multiplier 26 receives an input of the log-likelihood ratio LLRφ _{(bit_num)} based on the phase noise output from the phase noise LLR calculation unit 22, and also receives an input of the weighting coefficient Wφ of the log-likelihood ratio based on the phase noise output from the weighting coefficient control unit 24, multiplies the log-likelihood ratio LLRφ _{(bit_num)} based on the phase noise by the weighting coefficient Wφ of the log-likelihood ratio based on the phase noise, and outputs the multiplication result (i.e., Wφ×LLRφ _{(bit_num)} ).

The adder 27 receives the multiplication result (i.e., W _N ×LLR _{N(bit_num)} ) output from the first multiplier 25 and the multiplication result (i.e., Wφ×LLRφ _{(bit_num)} ) output from the second multiplier 26, adds these multiplication results according to the above equation 1, and outputs the sum, i.e., the log-likelihood ratio LLR _{bit_num} (= W _N ×LLR _{N(bit_num)} + Wφ×LLRφ _{(bit_num} )), for the (bit_num)th bit.

Then, the decoding unit 19 receives the log-likelihood ratio LLR _{bit_num} output from the LLR calculation unit 20, and performs low-density parity check decoding processing using the log-likelihood ratio LLR _{bit_num} according to, for example, the sum-product decoding method.

According to the LLR calculation unit 20 serving as a log-likelihood ratio calculation circuit according to the embodiment, the log-likelihood ratio LLR _{N(bit_num)} based on thermal noise and the log-likelihood ratio LLRφ _{(bit_num)} based on phase noise are taken into consideration of the degree of mutual priority between them (specifically, the log-likelihood ratio LLR _{N(bit_num)} is multiplied by the weighting factor W _N and the log-likelihood ratio LLRφ _{(bit_num)} is multiplied by the weighting factor Wφ), and then these are added to calculate the log-likelihood ratio LLR _{bit_num} for each bit. This allows the calculation of the log-likelihood ratio to take into consideration factors such as phase noise and distribution fluctuation due to fading in addition to the influence of thermal noise, making it possible to realize a low-density parity check that is robust against disturbances including phase noise.

The LLR calculation unit 20, which serves as a log-likelihood ratio calculation circuit according to the embodiment, also calculates the log-likelihood ratio by taking into account the probability distribution when the phase of the symbol rotates or fluctuates based on a normal distribution/Gaussian distribution due to phase noise or fading of the received signal. This makes it possible to improve the calculation accuracy of the log-likelihood ratio even in an environment where phase noise of the received signal is dominant, and ultimately makes it possible to improve the error correction capability of the low-density parity check.

Specifically, as shown in FIG. 10, the conventional method of calculating the log likelihood ratio is based on the assumption that the amplitude of the received signal fluctuates based on a normal distribution/Gaussian distribution centered on the ideal symbol point due to thermal noise in the received signal (see the dashed circle Cf in the figure). For this reason, if there is a fluctuation when the phase of the symbol rotates due to the phase noise of the received signal (see the circular arrow Fp in the figure), the signal Sf determined to have a high probability of being received will differ from the actual transmission signal Ss. In contrast, the LLR calculation unit 20 as the log likelihood ratio calculation circuit according to the embodiment is designed to take into account the distribution of fluctuations due to the phase noise of the received signal, that is, the log likelihood ratio is calculated taking into account the situation in which the symbol fluctuates on the envelope E due to the phase noise of the received signal, so that even if there is a fluctuation when the phase of the symbol rotates due to the phase noise of the received signal (see the circular arrow Fp in the figure), the signal determined to have a high probability of being received will match the actual transmission signal Ss, making it possible to improve the error correction ability of the low-density parity check.

An example of verifying the effectiveness of the log-likelihood ratio calculation circuit of this invention is described below.

In this verification example, the evaluation system shown in FIG. 11 was used for the purpose of comparing the calculation accuracy of the log likelihood ratio between a conventional method for calculating the log likelihood ratio and the log likelihood ratio calculation circuit according to the present invention in an environment where thermal noise and phase noise are present. The evaluation system in this verification example was configured as a system for comparing the calculation accuracy of the log likelihood ratio between a conventional method (reference number 34) that calculates the log likelihood ratio based on the distribution of thermal noise of the received signal after modulating a predetermined sample signal using the 16QAM method (reference number 31 in the figure) and adding phase noise and thermal noise (reference numbers 32 and 33), and the log likelihood ratio calculation circuit according to the present invention (reference number 20) that calculates the log likelihood ratio while also taking into account the phase noise of the received signal.

The evaluation conditions for this verification example were set as follows:
Modulation method: 16QAM method Carrier-to-noise ratio (CNR: Carrier-Noise Ratio): 15 to 25 dB
Phase noise level: -60 dBc (at 1 kHz offset)
Symbol interval S _d in QAM system: 2
Priority adjustment coefficient α: 0.125
Equivalent adjustment coefficient β: 100

The log-likelihood ratio takes a positive value when the probability that the transmitted bit is 0 is high, and takes a negative value when the probability that the transmitted bit is 1 is high. Using this, the actual transmitted bit is compared with the sign of the log-likelihood ratio calculated by the conventional method (reference number 34 in FIG. 9) to count the number of times the sign is incorrect (reference number 35), and the actual transmitted bit is compared with the sign of the log-likelihood ratio calculated by the log-likelihood ratio calculation circuit of this invention (reference number 20) to count the number of times the sign is incorrect (reference number 36).

As a result of the evaluation system shown in FIG. 11, it was confirmed that the log-likelihood ratio calculation circuit of the present invention had a smaller code error rate at each carrier-to-noise ratio, as shown in FIG. 12. It was also confirmed that the higher the carrier-to-noise ratio and the more dominant the phase noise, the greater the improvement rate achieved by the log-likelihood ratio calculation circuit of the present invention. From these results, it was confirmed that the code error rate is reduced by the log-likelihood ratio calculation circuit of the present invention compared to the conventional log-likelihood ratio calculation method.

The above describes an embodiment of the present invention, but the specific configuration is not limited to the above embodiment, and even if there are design changes within the scope of the invention that do not deviate from the gist of the invention, they are still included in the invention.

Specifically, in the above embodiment, the log-likelihood ratio calculation circuit according to the present invention is incorporated as an LLR calculation unit 20 in a wireless receiving device 1 whose schematic configuration is shown in FIG. 1, but the configuration of a wireless device in which the log-likelihood ratio calculation circuit according to the present invention can be incorporated is not limited to the wireless receiving device 1 whose schematic configuration is shown in FIG. 1, and the log-likelihood ratio calculation circuit according to the present invention may be incorporated in a wireless device of another configuration.

In addition, in the above embodiment, the log-likelihood ratio LLR _{N(bit_num)} based on thermal noise is calculated according to the above Formula 2A to Formula 2D, but the calculation formula/method of the log-likelihood ratio LLR _{N(bit_num)} based on thermal noise is not limited to Formula 2A to Formula 2D, and in the above embodiment, the log-likelihood ratio LLRφ _{(bit_num)} based on phase noise is calculated according to the above Formula 3 to Formula 6, but the calculation formula/method of the log-likelihood ratio LLRφ _{(bit_num)} based on phase noise is not limited to Formula 3 to Formula 6. That is, the gist of the present invention is to calculate the log-likelihood ratio LLR bit_num for each bit by taking into consideration the degree of mutual priority between the log-likelihood ratio LLR _{N (bit_num)} based on thermal noise and the log-likelihood ratio LLRφ _{(bit_num)} based on phase noise (specifically, by multiplying the log-likelihood ratio LLR N _{(bit_num)} by a weighting coefficient W _N and multiplying the log-likelihood ratio LLRφ _{(bit_num)} by a weighting coefficient Wφ) and adding these, and other configurations including the calculation formula/method of the log-likelihood ratio LLR _{N(bit_num)} based on thermal noise and the log-likelihood ratio LLRφ _{(bit_num)} based on phase noise are not limited to specific configurations.

REFERENCE SIGNS LIST 1 Radio receiving device 10 Antenna 11 Channel filter 12 Local oscillator 13 Mixer 14 Automatic gain control section 15 A/D converter 16 Digital quadrature detection section 17 Roll-off filter 18 Equalizer 19 Decoding section 20 LLR calculation section 21 Thermal noise LLR calculation section 22 Phase noise LLR calculation section 23 Thermal noise power estimation section 24 Weighting coefficient control section 25 First multiplier 26 Second multiplier 27 Adder

Claims (3)

a log likelihood ratio calculation circuit that calculates a log likelihood ratio by multiplying a log likelihood ratio based on thermal noise and a log likelihood ratio based on phase noise by a weighting factor and then adding the weighting factors,
The log-likelihood ratio based on the phase noise is
Of the 4-bit received signals modulated by quadrature amplitude modulation with a multi-level of 16 on the transmitting side,
For bits that are differentiated as 0 or 1 depending on the phase angle of the ideal symbol point,
using a phase angle of the received signal, calculating based on a probability density distribution that becomes 0 and a probability density distribution that becomes 1 according to the phase angle;
Regarding bits that can be divided into ideal symbol points that are distinguished as being 0 or 1 depending on the amplitude (called "amplitude-distinguishing symbol points") and ideal symbol points that are distinguished as being 0 or 1 depending on the phase angle (called "phase-distinguishing symbol points"),
determining whether the symbol point is the amplitude-distinguishing symbol point or the phase-distinguishing symbol point based on the amplitude of the received signal;
calculating the amplitude-distinguishing symbol point based on a probability density distribution that becomes 0 and a probability density distribution that becomes 1 according to the amplitude using the amplitude of the received signal;
calculating the phase-distinguishing symbol point based on a probability density distribution that becomes 0 and a probability density distribution that becomes 1 according to the phase angle using a phase angle of the received signal;
A log-likelihood ratio calculation circuit comprising: A log likelihood ratio based on phase noise is calculated for a first bit, which is a bit that is distinguished as being 0 or 1 depending on the phase angle of the ideal symbol point, according to the following Equation 1, and a log likelihood ratio based on phase noise is calculated for a third bit, according to the following Equation 3,
A log likelihood ratio based on phase noise is calculated for the second bit, which is a bit divided into the amplitude-distinguishing symbol point and the phase-distinguishing symbol point, according to the following Equation 2, and a log likelihood ratio based on phase noise is calculated for the fourth bit, according to the following Equation 4 (wherein Equation 2A and Equation 4A are used when _xR ≥ _Rthre1 or _xR ≤ _Rthre2 , and Equation 2B and Equation 4B are used when _Rthre2 < _xR < _Rthre1 ).
Here,
xφ: Phase angle of the received signal _xR : Amplitude of the received signal _φi : Phase angle of the ideal symbol point (where the subscript i of φ is 0, 1, 2, ..., 11)
R _i : Amplitude of the ideal symbol point (where the subscript i of R = 0, 1, 2)
2. The log-likelihood ratio calculation circuit according to claim 1 , wherein σ ² is a noise variance, R _thre1 is a first amplitude threshold, and R _thre2 is a second amplitude threshold. A log-likelihood ratio calculation circuit according to claim 1 or 2 ,
A radio receiving device comprising:

Images