技术领域technical field
本发明是关于通信系统的信号接收端，尤其是关于可适应性调整软信息(softinformation)的信号接收端及其信号处理方法。The present invention relates to a signal receiving end of a communication system, in particular to a signal receiving end capable of adaptively adjusting soft information and a signal processing method thereof.
背景技术Background technique
请参阅图1，其是已知通信系统的信号接收端的功能方块图。由发送端(未绘示)所传送的传输信号S经过通道110(具有通道效应H)传输及传输过程中噪声N的附加后(以加法器120表示)，信号接收端130接收到输入信号Y＝HS+N，通道估测单元131依据输入信号Y估测出输入信号Y所经历的通道效应H'，之后等化器132可依据通道效应H'来还原输入信号Y以得到调制信号S'＝Y/H'＝(HS+N)/H'。假设H'相当接近H，则调制信号S'可以近似为S'＝S+N'，N'为经过等化处理的噪声。之后软决策解调制电路(soft-decision demodulator)133将调制信号S'解调制，产生多个软信息SI。请参阅图2，其是软信息与硬信息的关系图。在硬决策(hard decision)中，如果调制信号S'的值大于临界值η，则调制信号S'将被判断为S1；而如果调制信号S'的值小于临界值η，则调制信号S'将被判断为S0。临界值η通常为S1与S0的平均值，即η＝(S0+S1)/2，所以如果调制信号S'刚好在临界值η附近，则很可能有误判的情形发生。另一方面，在软决策(soft decision)中，调制信号S'依据最大事后机率(maximum aposterior,MAP)法则来做判断，可以得出调制信号S'接近S1或接近S0的程度。举例来说，若调制信号S'位于y所示的位置，对硬决策而言，调制信号S'会被判定为S0；然而对软决策而言，调制信号S'被判定为S1的机率(P1)大于被判定为S0的机率(P0)的机率，所以最终会被判定为S1。依据最大事后机率法则，软决策会有较高的可靠度。然而软信息SI的准确度常常受到通道或噪声影响而降低，因此传统上会对软信息SI进行补偿。Please refer to FIG. 1 , which is a functional block diagram of a signal receiving end of a known communication system. After the transmission signal S transmitted by the transmitting end (not shown) is transmitted through the channel 110 (with the channel effect H) and the noise N is added during the transmission (represented by the adder 120), the signal receiving end 130 receives the input signal Y =HS+N, the channel estimation unit 131 estimates the channel effect H' experienced by the input signal Y according to the input signal Y, and then the equalizer 132 can restore the input signal Y according to the channel effect H' to obtain the modulated signal S'=Y/H'=(HS+N)/H'. Assuming that H' is quite close to H, the modulated signal S' can be approximated as S'=S+N', where N' is equalized noise. Afterwards, a soft-decision demodulator 133 demodulates the modulated signal S′ to generate a plurality of soft information SI . Please refer to Figure 2, which is a diagram of the relationship between soft information and hard information. In a hard decision, if the value of the modulation signal S' is greater than the critical value η, the modulation signal S' will be judged as S1 ; and if the value of the modulation signal S' is smaller than the critical value η, the modulation signal S ' will be judged as S0 . The critical value η is usually the average value of S1 and S0 , that is, η=(S0 +S1 )/2, so if the modulation signal S' is just around the critical value η, misjudgment is likely to occur. On the other hand, in the soft decision, the modulated signal S' is judged according to the maximum posterior probability (MAP) rule, and the degree to which the modulated signal S' is close to S1 or close to S0 can be obtained. For example, if the modulated signal S' is located at the position indicated by y, for the hard decision, the modulated signal S' will be judged as S0 ; however, for the soft decision, the modulated signal S' will be judged as S1 The probability (P1 ) is greater than the probability (P0 ) of being judged as S0 , so it will be judged as S1 in the end. According to the law of maximum ex post probability, soft decision-making will have higher reliability. However, the accuracy of the soft information SI is often reduced by channel or noise, so the soft information SI is traditionally compensated.
请继续参阅图1，软决策解调制电路133的输出耦接一个乘法器134，其对软信息SI乘上一个固定系数A，而产生调整后的软信息SI'，之后量化器135再依据其调整范围将调整后的软信息SI'量化而生成量化数据，最后解码器136再对量化数据进行解码而产生解码数据。然而以固定系数A来调整软信息SI有其缺点，如果当软信息SI够好时(代表经量化及解码后所产生的解码数据具有很低的位元错误率(bit error rate,BER))，将其乘上固定系数A反而可能造成位元错误率上升。因此有必要对已知的信号接收端130做改善，使量化器135可以得到更适当的软信息，以降低信号接收端130的位元错误率。Please continue to refer to FIG. 1, the output of the soft decision demodulation circuit 133 is coupled to a multiplier 134, which multiplies the soft information SI by a fixed coefficient A to generate adjusted soft information SI ', and then the quantizer 135 again The adjusted soft information SI ′ is quantized according to its adjustment range to generate quantized data, and finally the decoder 136 decodes the quantized data to generate decoded data. However, adjusting the soft information SI with a fixed coefficient A has its disadvantages. If the soft information SI is good enough (representing that the decoded data generated after quantization and decoding has a very low bit error rate (BER )), multiplying it by a fixed coefficient A may cause an increase in the bit error rate. Therefore, it is necessary to improve the known signal receiving end 130 so that the quantizer 135 can obtain more appropriate soft information to reduce the bit error rate of the signal receiving end 130 .
发明内容Contents of the invention
鉴于先前技术的不足，本发明的一目的在于提供一种可适应性调整软信息的信号接收端及其信号处理方法，以降低位元错误率。In view of the deficiencies of the prior art, an object of the present invention is to provide a signal receiving end capable of adaptively adjusting soft information and a signal processing method thereof, so as to reduce the bit error rate.
本发明揭示了一种通信系统的信号接收端，接收一调制信号，该信号接收端包含：一解调制电路，用来解调制该调制信号以产生多个封包，并产生对应于每一封包的多个软信息；一软信息调整电路，耦接该解调制电路，用来依据对应于该每一封包的这些软信息的一分布情形调整这些软信息；一量化器，耦接该软信息调整电路，用来量化这些调整后的软信息以产生多个数据；以及一解码器，耦接该量化器，用来解码这些数据。The invention discloses a signal receiving end of a communication system, which receives a modulated signal, and the signal receiving end includes: a demodulation circuit, which is used to demodulate the modulated signal to generate a plurality of packets, and generate a signal corresponding to each packet A plurality of soft information; a soft information adjustment circuit, coupled to the demodulation circuit, used to adjust these soft information according to a distribution of the soft information corresponding to each packet; a quantizer, coupled to the soft information adjustment A circuit, used to quantize the adjusted soft information to generate a plurality of data; and a decoder, coupled to the quantizer, used to decode the data.
上述的软信息调整电路包含：一第一计算电路，耦接该解调制电路，用来依据这些软信息的该分布情形计算得到一索引值；一查找单元，耦接该索引值计算电路与该查找表，用来依据该索引值查找该查找表得到一特征值；一第二计算电路，耦接该查找单元，用来依据该特征值及这些软信息的分布情形产生该缩放系数；以及一乘法单元，耦接该第二计算电路及该解调制电路，用来将这些软信息乘上该缩放系数以产生这些调整后的软信息。The above-mentioned soft information adjustment circuit includes: a first calculation circuit, coupled to the demodulation circuit, used to calculate an index value according to the distribution of the soft information; a search unit, coupled to the index value calculation circuit and the A lookup table, used to look up the lookup table according to the index value to obtain a feature value; a second calculation circuit, coupled to the lookup unit, used to generate the scaling factor according to the feature value and the distribution of the soft information; and a The multiplication unit, coupled to the second calculation circuit and the demodulation circuit, is used for multiplying the soft information by the scaling factor to generate the adjusted soft information.
本发明另揭示了一种信号处理方法，应用于通信系统的信号接收端，包含：解调制一调制信号以产生多个封包，并产生对应于每一封包的多个软信息；依据对应于该每一封包的这些软信息的一分布情形调整这些软信息；量化这些调整后的软信息以产生多个数据；以及解码这些数据。The present invention also discloses a signal processing method applied to a signal receiving end of a communication system, comprising: demodulating a modulated signal to generate a plurality of packets, and generating a plurality of soft information corresponding to each packet; according to the A distribution of the soft information for each packet adjusts the soft information; quantizes the adjusted soft information to generate a plurality of data; and decodes the data.
上述的该依据这些软信息的分布情形查找该查找表以产生该缩放系数的步骤包含：依据这些软信息的该分布情形计算得到一索引值；依据该索引值查找该查找表以得到一特征值；以及依据该特征值及这些软信息产生该缩放系数。The step of searching the lookup table according to the distribution of the soft information to generate the scaling factor includes: calculating an index value according to the distribution of the soft information; searching the lookup table according to the index value to obtain a feature value ; and generating the scaling factor according to the feature value and the soft information.
本发明的可适应性调整软信息的信号接收端及其信号处理方法可以依据软信息的分布情形来适当地调整软信息，经过调整的软信息可以更能有效地被下一级的量化器量化，因此最后产生的解码数据也更正确。相较于已知的通信系统的信号接收端，本发明更可以依据软信息的实际分布情形，动态调整软信息，使软信息得到较佳的补偿效果。已知使用固定系数调整软信息无法依实际的信号状况作动态调整，因此有时甚至造成反效果，导致补偿后的软信息反而增加位元错误率。The adaptively adjustable soft information signal receiving end and its signal processing method of the present invention can properly adjust the soft information according to the distribution of the soft information, and the adjusted soft information can be more effectively quantized by the next-level quantizer , so the resulting decoded data is also more correct. Compared with the signal receiving end of the known communication system, the present invention can dynamically adjust the soft information according to the actual distribution of the soft information, so that the soft information can obtain a better compensation effect. It is known that using a fixed coefficient to adjust the soft information cannot be dynamically adjusted according to the actual signal conditions, so sometimes it even has the opposite effect, causing the compensated soft information to increase the bit error rate instead.
有关本发明的特征、实作与功效，兹配合附图作实施例详细说明如下。The features, implementation and effects of the present invention are described in detail as follows with reference to the accompanying drawings.
附图说明Description of drawings
图1为已知通信系统的信号接收端的功能方块图；Fig. 1 is the functional block diagram of the signal receiving end of known communication system;
图2为软信息与硬信息的关系图；Figure 2 is a relationship diagram between soft information and hard information;
图3为本发明通信系统的信号接收端的功能方块图；Fig. 3 is a functional block diagram of the signal receiving end of the communication system of the present invention;
图4为软信息的分布情形的示意图；FIG. 4 is a schematic diagram of a distribution situation of soft information;
图5为本发明软信息调整电路的一实施例的功能方块图；5 is a functional block diagram of an embodiment of the soft information adjustment circuit of the present invention;
图6为本发明当调制信号为二进制相移键控调制的索引值I及封包位元错误率与SNR的关系图；Fig. 6 is the relationship diagram of index value I and packet bit error rate and SNR when modulated signal is binary phase-shift keying modulation of the present invention;
图7为本发明当调制信号为四相相移键控调制的索引值I及封包位元错误率与SNR的关系图；Fig. 7 is when the modulation signal of the present invention is the index value I of quadrature phase-shift keying modulation and the relationship diagram of packet bit error rate and SNR;
图8为本发明用来建立查找表的一实施方式的电路图；Fig. 8 is the circuit diagram of an embodiment that the present invention is used to establish look-up table;
图9为本发明对应BPSK调制机制的查找表的其中一实施例；Fig. 9 is one embodiment of the look-up table corresponding to the BPSK modulation mechanism of the present invention;
图10为本发明对应QPSK调制机制的查找表的其中一实施例；Fig. 10 is one embodiment of the look-up table corresponding to the QPSK modulation mechanism of the present invention;
图11为本发明对应64QAM调制机制的查找表的其中一实施例；FIG. 11 is one embodiment of the look-up table corresponding to the 64QAM modulation mechanism of the present invention;
图12为BPSK调制机制下使用已知的固定系数与使用本发明的最佳系数的位元错误率与SNR的关系图；Fig. 12 is a relationship diagram between the bit error rate and SNR using known fixed coefficients and using the best coefficients of the present invention under the BPSK modulation mechanism;
图13为QPSK调制机制下使用已知的固定系数与使用本发明的最佳系数的位元错误率与SNR的关系图；Fig. 13 is a relationship diagram between bit error rate and SNR using known fixed coefficients and using the best coefficients of the present invention under the QPSK modulation mechanism;
图14为64QAM调制机制下使用已知的固定系数与使用本发明的最佳系数的位元错误率与SNR的关系图；Fig. 14 is a relationship diagram between the bit error rate and the SNR using known fixed coefficients and using the best coefficients of the present invention under the 64QAM modulation mechanism;
图15为本发明的信号处理方法的一实施例的流程图；以及Fig. 15 is a flowchart of an embodiment of the signal processing method of the present invention; and
图16为本发明依据软信息的分布情形动态调整软信息的细部流程图。Fig. 16 is a detailed flow chart of the present invention to dynamically adjust soft information according to the distribution of soft information.
符号说明Symbol Description
110：通道110: channel
120：加法器120: Adder
130、310：信号接收端130, 310: signal receiving end
131：通道估测单元131: channel estimation unit
132：等化器132: Equalizer
133：软决策解调制电路133: Soft decision demodulation circuit
134、820、540：乘法器134, 820, 540: multiplier
135：量化器135: Quantizer
136：解码器136: Decoder
312：软信息调整电路312: Soft information adjustment circuit
410、420、430：分布态样410, 420, 430: Distribution pattern
510、530、830：计算电路510, 530, 830: computing circuits
520：查找表单元520: Lookup table unit
810：通道模型810: Channel model
840：选择器840: selector
S1510～S1530、S1610～S1640：步骤S1510～S1530, S1610～S1640: steps
SI、SI’：软信息SI , SI ': soft information
具体实施方式Detailed ways
以下说明内容的技术用语系参照本技术领域的习惯用语，如本说明书对部分用语有加以说明或定义，该部分用语的解释是以本说明书的说明或定义为准。The technical terms in the following explanations refer to the customary terms in this technical field. If some terms are explained or defined in this specification, the explanations of these terms shall be based on the descriptions or definitions in this specification.
本发明的揭示内容包含可适应性调整软信息的信号接收端及其信号处理方法，能依据软信息的分布情形适应性地调整软信息。其中，该装置与方法可应用于利用正交振幅解调制(quadrature amplitude modulation,QAM)的通信系统。The disclosed content of the present invention includes a signal receiving end capable of adaptively adjusting soft information and a signal processing method thereof, which can adaptively adjust soft information according to the distribution of soft information. Wherein, the device and method can be applied to a communication system using quadrature amplitude modulation (QAM).
请参阅图3，其是本发明通信系统的信号接收端的功能方块图。本发明的信号接收端310包含但不限于通道估测单元131、等化器132、软决策解调制电路133、软信息调整电路312、量化器135以及解码器136，图3的功能方块图仅绘示与本发明直接相关的元件，但依实际情况，本发明的信号接收端310可能包含其他元件，例如解交错(de-interleaving)电路及降频电路等。信号接收端310中与图1的信号接收端130具有相同名称及标号的元件具有相同的功能，故不再赘述。本发明的软信息调整电路312可依据软信息的分布情形对其做调整，使调整后的软信息更适于被量化器135量化，因此最终解码器136所产生的解码数据有较低的位元错误率。Please refer to FIG. 3 , which is a functional block diagram of the signal receiving end of the communication system of the present invention. The signal receiving end 310 of the present invention includes but is not limited to a channel estimation unit 131, an equalizer 132, a soft decision demodulation circuit 133, a soft information adjustment circuit 312, a quantizer 135, and a decoder 136. The functional block diagram of FIG. 3 is only Components directly related to the present invention are shown, but according to actual conditions, the signal receiving end 310 of the present invention may include other components, such as de-interleaving circuits and down-frequency circuits. Components in the signal receiving end 310 with the same names and labels as those of the signal receiving end 130 in FIG. 1 have the same functions, and thus will not be described again. The soft information adjustment circuit 312 of the present invention can adjust it according to the distribution of the soft information, so that the adjusted soft information is more suitable for being quantized by the quantizer 135, so the decoded data generated by the final decoder 136 has a lower bit Meta error rate.
请参阅图4，其是软信息的分布情形的示意图。传送端在传送数据时通常会以一个封包(packet)为传输单位，因此接收端在处理信号时也以一个封包为单位做处理。图4的纵轴代表数据个数，而横轴代表软信息，也称为LLR(Log Likelihood Ratio，对数似然比)，是由最大事后机率所得到，其数学式如方程式(1)所示：Please refer to FIG. 4 , which is a schematic diagram of the distribution of soft information. The transmitting end usually uses a packet as a transmission unit when transmitting data, so the receiving end also processes a signal in a packet as a unit. The vertical axis in Figure 4 represents the number of data, while the horizontal axis represents soft information, also known as LLR (Log Likelihood Ratio, log likelihood ratio), which is obtained by the maximum ex post probability, and its mathematical formula is as shown in equation (1). Show:
当P(S|S＝S1)>P(S|S＝S0)(代表信号S被判定为S1的机率大于被判定为S0的机率)，LLR(S)>0；当P(S|S＝S1)<P(S|S＝S0)(代表信号S被判定为S1的机率小于被判定为S0的机率)，LLR(S)<0。图4的软信息分布图包含三种分布态样410、420及430，分别对应信号噪声比(signal-to-noise ratio，以下简称SNR)为-5dB、0dB及5dB，每个分布态样皆对应一个封包，以多个长条图分别表示某个LLR值所具有的数据个数。在本例中，一个封包具有10万笔数据，而且图4是描绘软信息的绝对值。从图中可以看出，当SNR愈大时，一个封包的软信息的分布愈广，也就是分布态样愈宽。分布态样的宽或窄将影响量化器135的量化结果，如果分布太宽，以致于软信息的最大值(或最小值)超出量化器135量化范围的上限(或下限)，例如软信息的分布范围为-10～+10，而量化范围为-8～+7，则小于-8及大于+7的软信息皆被量化为同样的量化数据，导致量化的解析度下降；另一方面，如果分布太窄，软信息的分布集中在量化范围的一小段，例如量化范围为-8～+7，而软信息的分布范围为-3～+2，则量化数据的准确度便会下降。因此本发明提出可以依据软信息的实际分布情形动态调整软信息的软信息调整电路312。When P(S|S=S1 )>P(S|S=S0 ) (representing that the probability of signal S being judged as S1 is greater than the probability of being judged as S0 ), LLR(S)>0; when P (S|S=S1 )<P(S|S=S0 ) (representing that the probability of signal S being judged as S1 is smaller than the probability of being judged as S0 ), LLR(S)<0. The soft information distribution diagram in Figure 4 includes three distribution patterns 410, 420, and 430, corresponding to the signal-to-noise ratio (SNR) of -5dB, 0dB, and 5dB respectively, and each distribution pattern is Corresponding to one packet, multiple bar graphs are used to represent the number of data contained in a certain LLR value. In this example, one package has 100,000 data, and Fig. 4 depicts the absolute value of soft information. It can be seen from the figure that when the SNR is larger, the distribution of soft information of a packet is wider, that is, the distribution pattern is wider. The width or narrowness of the distribution pattern will affect the quantization result of the quantizer 135, if the distribution is too wide, so that the maximum value (or minimum value) of the soft information exceeds the upper limit (or lower limit) of the quantization range of the quantizer 135, such as the soft information The distribution range is -10 to +10, and the quantization range is -8 to +7, so the soft information less than -8 and greater than +7 will be quantized into the same quantization data, resulting in a decrease in quantization resolution; on the other hand, If the distribution is too narrow, the distribution of soft information is concentrated in a small part of the quantization range, for example, the quantization range is -8 to +7, and the distribution range of soft information is -3 to +2, then the accuracy of the quantization data will decrease. Therefore, the present invention proposes a soft information adjustment circuit 312 that can dynamically adjust the soft information according to the actual distribution of the soft information.
请参阅图5，其是本发明软信息调整电路312的一实施例的功能方块图。软信息调整电路312包含计算电路510、查找表单元520以及计算电路530。计算电路510耦接软决策解调制电路133，用来依据软决算值SI的分布情形而计算出索引值I。更详细地说，为了要得到软信息的分布情形，计算电路510依据方程式(2)计算索引值I：Please refer to FIG. 5 , which is a functional block diagram of an embodiment of the soft information adjusting circuit 312 of the present invention. The soft information adjustment circuit 312 includes a calculation circuit 510 , a lookup table unit 520 and a calculation circuit 530 . The calculation circuit 510 is coupled to the soft decision demodulation circuit 133 for calculating the index value I according to the distribution of the soft decision value SI . In more detail, in order to obtain the distribution of soft information, the calculation circuit 510 calculates the index value I according to equation (2):
E[|SI|]代表软信息SI的绝对值的平均，Var[|SI|]是软信息SI的绝对值的变异量，事实上Var[|SI|]＝E[(|SI|-E[|SI|])2]。因此索引值I可以反应软信息SI的分布形情，其与分布态样的中心点(即E[|SI|])及分布范围的宽窄有关。请参阅图6，其是本发明当调制信号为二进制相移键控(binary phase shift keying,以下简称BPSK)调制的索引值I及封包位元错误率(packet bit error rate，以下简称PBER)与SNR的关系图。PBER为数据错误的封包个数与所有封包个数的比值，而当数据封包中有任何一笔数据错误时，该封包视为数据错误的封包，只有在封包中的所有数据皆正确时，该封包才视为数据正确的封包。因此PBER为0代表数据的准确率为100％，而为1代表数据的准确率极差。由图中可以发现，在BPSK的调制机制下，索引值I随着SNR的增加而递减，而PBER在高SNR时(此例为SNR>-0.5dB)为0，在低SNR时(此例为SNR<-1.2dB)为1。在此实施例中，当PBER为0时代表原本的软信息SI就已经够好，无需进一步调整，因此本发明可针对PBER不为0的状况(此例为SNR<-0.5dB)进行软信息SI调整，然而当PBER为1时不仅代表原本的软信息SI不佳，更代表其错误率已经太高，即使对其做调整效果亦不大，所以实作上，本发明于此实施例中更专注于调整PBER为0.1～0.9之间的情况，也就是针对SNR介于-0.6dB与-1.0dB之间时针对软信息SI做调整，在此实施例中，SNR介于-0.6dB与-1.0dB之间对应的索引值I是约略介于0.5124与0.534之间。请参阅图7，其是本发明当调制信号为四相相移键控(quadrature phase shift keying,以下简称QPSK)调制的索引值I及封包位元错误率与SNR的关系图。从图中可以发现，当SNR增加时索引值I亦呈现递减的状况，而且PBER在高SNR时(此例为SNR>2.3dB)为0，在低SNR时(此例为SNR<1.7dB)为1。同理，本发明在这个实施例中可以更专注于调整PBER为0.1～0.9之间的情况，也就是针对SNR介于1.9dB与2.3dB之间时针对软信息SI做调整，在此实施例中，SNR介于1.9dB与2.3dB之间对应的索引值I是约略介于0.5339与0.5106之间。E[|SI |] represents the average of the absolute value of soft information SI , Var[|SI |] is the variation of the absolute value of soft information SI , in fact Var[|SI |]=E[( |SI |-E[|SI |])2 ]. Therefore, the index value I can reflect the distribution of soft information SI , which is related to the center point of the distribution pattern (ie, E[|SI |]) and the width of the distribution range. Please refer to FIG. 6, which shows the index value I and the packet bit error rate (packet bit error rate, hereinafter referred to as PBER) and SNR plot. PBER is the ratio of the number of packets with data errors to the number of all packets. When there is any data error in a data packet, the packet is regarded as a packet with data errors. Only when all the data in the packet is correct, the PBER The packet is regarded as a packet with correct data. Therefore, a PBER of 0 means that the accuracy of the data is 100%, and a PBER of 1 means that the accuracy of the data is extremely poor. It can be found from the figure that under the modulation mechanism of BPSK, the index value I decreases with the increase of SNR, while PBER is 0 at high SNR (in this example, SNR>-0.5dB), and at low SNR (in this example 1 for SNR<-1.2dB). In this embodiment, when the PBER is 0, it means that the original soft information S1 is good enough and no further adjustment is needed. Therefore, the present invention can perform soft information for the situation where the PBER is not 0 (in this example, SNR<-0.5dB). Information SI adjustment, but when PBER is 1, it not only means that the original soft information SI is not good, but also means that its error rate is too high, even if it is adjusted, the effect is not great, so in practice, the present invention is here In the embodiment, more focus is placed on adjusting the PBER between 0.1 and 0.9, that is, when the SNR is between -0.6dB and -1.0dB, the soft information SI is adjusted. In this embodiment, the SNR is between The corresponding index value I between -0.6dB and -1.0dB is approximately between 0.5124 and 0.534. Please refer to FIG. 7 , which is a graph showing the relationship between index value I, packet error rate and SNR when the modulation signal is quadrature phase shift keying (QPSK) modulation in the present invention. It can be seen from the figure that the index value I also decreases when the SNR increases, and the PBER is 0 when the SNR is high (SNR>2.3dB in this example), and is 0 when the SNR is low (SNR<1.7dB in this example) is 1. Similarly, in this embodiment, the present invention can focus more on adjusting the situation where the PBER is between 0.1 and 0.9, that is, when the SNR is between 1.9dB and 2.3dB, the soft information SI is adjusted. In the example, the index value I corresponding to the SNR between 1.9dB and 2.3dB is approximately between 0.5339 and 0.5106.
请继续参阅图5，计算电路510计算出索引值I后，查找表单元520依据索引值I查找一查找表以得到特征值α，该查找表可以储存于查找表单元520的内部或独立于查找表单元520外的储存空间(未绘示)。查找表事先依据不同的通道以及调制机制，以实测或模拟的方式建立。请参阅图8，其是本发明用来建立查找表的一实施方式的电路图。首先决定通道模型810的种类，例如常见的衰落通道(fading channel，或称为多路径衰落通道(multipathfading channel))及平坦通道(flat channel)(不以此为限)，再决定调制机制，例如BPSK、QPSK、16QAM及64QAM等(不以此为限)。通道种类及调制机制确定后，便可藉由改变SNR来产生对应不同SNR的软信息。软信息之后被多个乘法器820分别乘上不同的特征值(α1、α2、…、αn)，再由计算电路830计算个别的PBER，最后再由比较器840找出最小的PBER，而对应该最小的PBER的特征值α即为该SNR下的最佳特征值。藉由不同通道模型810的种类、不同调制机制、以及不同SNR，即可估算出对应不同条件下的最佳特征值并存成前述的查找表。请参阅图9，其是本发明对应BPSK调制机制的查找表的其中一实施例。如前揭所述，在BPSK的调制机制下，当SNR落于-0.6dB～-1.0dB才需要调整软信息，因此在建立图9的查找表时，在上述的范围内SNR每变化0.1dB即测量一次特征值，所以共得到5个特征值，然而可以缩小SNR的测量间隔，以提升查找表的准确度。得到SNR与特征值α的关系后，可以进一步从图6找出SNR与索引值I的关系，因此最后查找表单元520可以仅依据索引值I来找出对应的特征值α。同理，其他的调制机制亦有不同的查找表，图10为本发明对应QPSK调制机制的查找表的其中一实施例，如前揭所述，在QPSK的调制机制下，当SNR落于1.9dB～-2.3dB才需要调整软信息，因此对于QPSK调制机制将只会对SNR介于1.9dB与2.3dB之间的部分建立查找表；图11为本发明对应64QAM调制机制的查找表的其中一实施例。请注意，事先建立查找表可以增加查表的效率，并且精简电路的设计，然而图8的电路亦可整合于软信息调整电路312中，以便及时且依需求进行模拟并提供查找表单元520所需的查找表。Please continue to refer to FIG. 5, after the calculation circuit 510 calculates the index value I, the lookup table unit 520 searches a lookup table according to the index value I to obtain the feature value α, and the lookup table can be stored in the lookup table unit 520 or independent of the lookup table. A storage space (not shown) outside the table unit 520 . The look-up table is established in advance by actual measurement or simulation according to different channels and modulation mechanisms. Please refer to FIG. 8 , which is a circuit diagram of an embodiment of the present invention for establishing a look-up table. First determine the type of channel model 810, such as common fading channel (fading channel, or multipath fading channel (multipathfading channel)) and flat channel (flat channel) (not limited to this), and then determine the modulation mechanism, such as BPSK, QPSK, 16QAM and 64QAM, etc. (not limited to this). After the channel type and modulation mechanism are determined, soft information corresponding to different SNRs can be generated by changing the SNR. The soft information is then multiplied by different eigenvalues (α1 , α2 , ..., αn ) by multiple multipliers 820, and then the individual PBER is calculated by the calculation circuit 830, and finally the minimum PBER is found by the comparator 840 , and the eigenvalue α corresponding to the smallest PBER is the best eigenvalue under the SNR. With different types of channel models 810 , different modulation schemes, and different SNRs, the best feature values corresponding to different conditions can be estimated and stored in the aforementioned look-up table. Please refer to FIG. 9 , which is one embodiment of the look-up table corresponding to the BPSK modulation scheme of the present invention. As mentioned above, under the BPSK modulation mechanism, the soft information needs to be adjusted when the SNR falls between -0.6dB ~ -1.0dB. Therefore, when the look-up table in Figure 9 is established, the SNR changes by 0.1dB within the above range That is, the eigenvalue is measured once, so a total of 5 eigenvalues are obtained. However, the measurement interval of the SNR can be reduced to improve the accuracy of the lookup table. After obtaining the relationship between SNR and eigenvalue α, the relationship between SNR and index value I can be further found from FIG. Similarly, other modulation schemes also have different look-up tables. FIG. 10 is one embodiment of the look-up table corresponding to the QPSK modulation scheme of the present invention. As mentioned above, under the QPSK modulation scheme, when the SNR falls below 1.9 dB～-2.3dB just need to adjust the soft information, so will only set up the look-up table to the part between 1.9dB and 2.3dB for the QPSK modulation mechanism; Fig. 11 is wherein of the look-up table of the corresponding 64QAM modulation mechanism an embodiment. Please note that establishing the look-up table in advance can increase the efficiency of the look-up table and simplify the design of the circuit. However, the circuit in FIG. required lookup table.
请继续参阅图5，利用查找表单元520得到特征值α后，计算电路530再依据软信息SI对特征值α做计算而得到缩放系数f。缩放系数f为特征值α除以软信息SI的绝对值的平均，可以用下方的方程式(3)表示：Please continue to refer to FIG. 5 , after using the lookup table unit 520 to obtain the feature value α, the calculation circuit 530 calculates the feature value α according to the soft information SI to obtain the scaling factor f. The scaling factor f is the average of the feature value α divided by the absolute value of the soft information SI , which can be expressed by the following equation (3):
计算电路530可以直接取用计算电路510于计算的过程中所产生的E[|SI|]，也可以直接取软信息SI作运算。最后乘法器540将软信息SI与缩放系数f相乘而得到调整后的软信息SI'，最终软信息调整电路312便将调整后的软信息SI'输出给量化器135。以QPSK的一个实际例子来说明，当软信息SI的分布态样的Var[|SI|]＝8.6087及E[|SI|]＝4.0703，依据方程式(2)可以得到索引值I＝0.5196，且依据图10的查找表，由于索引值I＝0.5196小于0.5227，所以选择特征值α为4.7234，并依据方程式(3)得到其缩放系数f＝4.7234/4.0703＝1.1695。The calculation circuit 530 can directly use the E[|SI |] generated by the calculation circuit 510 during the calculation process, or can directly use the soft information SI for calculation. Finally, the multiplier 540 multiplies the soft information SI by the scaling factor f to obtain adjusted soft information SI ′, and finally the soft information adjustment circuit 312 outputs the adjusted soft information SI ′ to the quantizer 135 . Taking a practical example of QPSK to illustrate, when Var[|SI |]=8.6087 and E[|SI |]=4.0703 of the distribution pattern of soft information SI , according to equation (2), the index value I= 0.5196, and according to the lookup table in Figure 10, since the index value I=0.5196 is less than 0.5227, the feature value α is selected as 4.7234, and its scaling factor f=4.7234/4.0703=1.1695 is obtained according to equation (3).
请参阅图12，其是BPSK调制机制下使用已知的固定缩放系数与使用本发明的最佳系数(也就是依据软信息SI的分布情形所得到的缩放系数f)的位元错误率与SNR的关系图。可以发现，在软信息调整电路312的调整范围内，使用本发明的调整机制可以有效降低数据的位元错误率。对BPSK调制机制而言，平均来说，本发明相较于已知的方法约改善了0.07dB。图13及图14分别显示QPSK及64QAM的调制机制下，使用已知的固定系数与使用本发明的最佳系数的位元错误率与SNR的关系图。同样的，在软信息调整电路312的调整范围内，本发明可以有效降低位元错误率。Please refer to Fig. 12, which is the bit error rate and the relationship between the known fixed scaling coefficient and the optimal coefficient of the present invention (that is, the scaling coefficient f obtained according to the distribution of the soft information S1 ) under the BPSK modulation mechanism SNR plot. It can be found that within the adjustment range of the soft information adjustment circuit 312, the bit error rate of data can be effectively reduced by using the adjustment mechanism of the present invention. For the BPSK modulation scheme, on average, the present invention provides an improvement of about 0.07 dB compared to known methods. FIG. 13 and FIG. 14 respectively show the relationship between BER and SNR using known fixed coefficients and using the optimal coefficients of the present invention under QPSK and 64QAM modulation schemes. Likewise, within the adjustment range of the soft information adjustment circuit 312, the present invention can effectively reduce the bit error rate.
请参阅图15，其是本发明的信号处理方法的一实施例的流程图。除前述的通信系统的信号接收端外，本发明亦相对应地揭示了一种信号处理方法，应用于通信系统的信号接收端，能依据软信息SI的分布情形对其做动态调整，以更加适应、之后量化程序的特定量化范围。本方法由前揭信号接收端310或其等效装置来执行。如图15所示，本发明信号处理方法的一实施例包含下列步骤：Please refer to FIG. 15 , which is a flowchart of an embodiment of the signal processing method of the present invention. In addition to the above-mentioned signal receiving end of the communication system, the present invention also correspondingly discloses a signal processing method, which is applied to the signal receiving end of the communication system, and can be dynamically adjusted according to the distribution of the soft informationSI , so as to More adapted, specific quantization ranges for subsequent quantization procedures. The method is implemented by the aforementioned signal receiving end 310 or an equivalent device thereof. As shown in Figure 15, an embodiment of the signal processing method of the present invention comprises the following steps:
步骤S1510：解调制调制信号，并且产生多个软信息SI。通信系统的信号接收端所接收到的信号通常为经过调制的信号，此步骤利用以软决策为基础的解调制方法来解调制信号。解调制后可得到多个软信息SI；Step S1510: Demodulate the modulated signal, and generate a plurality of soft information SI . The signal received by the signal receiving end of the communication system is usually a modulated signal. In this step, a soft decision-based demodulation method is used to demodulate the signal. Multiple soft information SI can be obtained after demodulation;
步骤S1520：依据这些软信息SI的分布情形动态调整这些软信息SI。步骤S1510所产生的软信息SI可能受到通道及噪声等影响而变得较不理想，造成位元错误率太高，进而使整个封包数据错误。为了降低位元错误率，本发明依据软信息SI的分布情形动态调整软信息SI，而非以固定的系数来做调整，如此可以依据信号所实际经历的通道及噪声等环境状态(影响软信息SI的分布)，来动态调整软信息SI，因此调整后的软信息SI'可以更符合之后量化步骤的特定量化范围，以进一步降低位元错误率；以及Step S1520: Dynamically adjust the soft information SI according to the distribution of the soft information SI . The soft information SI generated in step S1510 may be affected by channel and noise and become less ideal, resulting in too high a bit error rate, and thus causing errors in the entire packet data. In order to reduce the bit error rate, the present invention dynamically adjusts the soft information SI according to the distribution of the soft information SI , instead of adjusting with a fixed coefficient, so that it can be based on the environmental conditions such as channels and noises actually experienced by the signal (affecting distribution of soft information SI ) to dynamically adjust soft information SI , so the adjusted soft information SI ' can be more in line with the specific quantization range of subsequent quantization steps to further reduce the bit error rate; and
步骤S1530：量化这些调整后的软信息SI'以产生多个量化数据。上一步骤所产生的调整后的软信息SI'已经更符合本步骤的量化范围，所以本步骤可以得到更精确的量化数据，有助于之后的解码步骤(未绘示)解码出位元错误率更低的解码数据。Step S1530: Quantize the adjusted soft information SI ′ to generate a plurality of quantized data. The adjusted soft information SI ' produced in the previous step is more in line with the quantization range of this step, so this step can obtain more accurate quantization data, which will help the subsequent decoding step (not shown) to decode bits Decoded data with lower error rate.
请注意，上述的步骤仅描述与本发明最相关的部分，一般而言，通信系统的信号接收端的信号处理方法可能包含其他的步骤，如图3所示的由通道估测单元131所执行的通道估测步骤，以及由等化器132所执行的等化步骤，或是其他例如解交错等步骤。再者，上述的步骤S1520还包含细部的子步骤，请参阅图16，其是本发明依据软信息的分布情形动态调整软信息的细部流程图，包含以下步骤：Please note that the above-mentioned steps only describe the most relevant part of the present invention. Generally speaking, the signal processing method at the signal receiving end of the communication system may include other steps, as shown in FIG. 3 performed by the channel estimation unit 131 The channel estimation step, and the equalization step performed by the equalizer 132, or other steps such as de-interleaving. Furthermore, the above step S1520 also includes detailed sub-steps, please refer to Figure 16, which is a detailed flow chart of the present invention for dynamically adjusting soft information according to the distribution of soft information, including the following steps:
步骤S1610：依据这些软信息SI的分布情形计算得到索引值I。一个封包的软信息的分布情形如图4所示，其分布态样随着SNR的不同而变化。而分布态样的宽或窄将影响量化的结果，所以本步骤依据软信息SI的分布情形产生索引值I，也就是说，索引值I可以反应软信息SI的分布情形。而索引值I与软信息SI的关系如方程式(2)所示；Step S1610: Calculate the index value I according to the distribution of these soft information SI. The distribution of the soft information of a packet is shown in Figure 4, and its distribution pattern varies with the SNR. The width or narrowness of the distribution pattern will affect the quantization result, so this step generates the index value I according to the distribution of the soft information SI , that is, the index value I can reflect the distribution of the soft information SI. The relationship between the index value I and the soft information SI is shown in equation (2);
步骤S1620：依据该索引值I以查表方式得到特征值α。查找表是依据不同的SNR值所对应的特征值α而建立，不同的通道模型及信号的调制机制将得到不同的查找表。建立查找表的细节已揭示于图8及其相关的描述，故不再赘述。而且如图6及图7所示，在不同的调制机制下，索引值I与SNR有一定的对应关系，因此藉由查表可以利用索引值I找到特征值α，也就是说特征值α是依据软信息SI的分布情形而查表得到；Step S1620: According to the index value I, the feature value α is obtained by looking up a table. The lookup table is established based on the eigenvalue α corresponding to different SNR values, and different channel models and signal modulation mechanisms will result in different lookup tables. The details of establishing the lookup table have been disclosed in FIG. 8 and related descriptions, so details are not repeated here. Moreover, as shown in Figures 6 and 7, under different modulation schemes, the index value I has a certain correspondence with the SNR, so the index value I can be used to find the eigenvalue α by looking up the table, that is to say, the eigenvalue α is According to the distribution situation of the soft information SI , it is obtained by looking up the table;
步骤S1630：依据特征值α及这些软信息SI的分布情形产生缩放系数f。缩放系数f可以由方程式(3)计算得到，也就是说缩放系数f是特征值α除以软信息SI的绝对值的平均；以及Step S1630: Generate a scaling factor f according to the feature value α and the distribution of the soft information SI . The scaling factor f can be calculated by equation (3), that is to say, the scaling factor f is the average of the absolute value of the feature value α divided by the soft information SI ; and
步骤S1640：将这些软信息SI乘上缩放系数f以产生这些调整后的软信息SI'。调整后的软信息SI'将有更适于量化操作的分布范围，有助进一步降低解码后的解码数据的位元错误率。Step S1640: Multiply the soft information SI by the scaling factor f to generate the adjusted soft information SI '. The adjusted soft information SI ′ will have a distribution range more suitable for the quantization operation, which helps to further reduce the bit error rate of the decoded decoded data.
综上所述，本发明依据软信息的分布情形来即时调整软信息，使调整时能依据软信息的实际状况做出适应性的调整，而非以固定的系数来做整，因此调整后的软信息更符合之后的量化程序所采用的特定量化范围，以进一步降低位元错误率。To sum up, the present invention adjusts the soft information in real time according to the distribution of the soft information, so that adaptive adjustments can be made according to the actual situation of the soft information during the adjustment, instead of adjusting with a fixed coefficient. Therefore, the adjusted The soft information is more consistent with the specific quantization range used by the subsequent quantization process to further reduce the bit error rate.
本技术领域具有通常知识者可藉由图3及图5的装置发明的揭露内容来了解图15及图16的方法发明的实施细节与变化。再者，前揭实施例虽以正交振幅调制为例，然此并非对本发明的限制，本技术领域人士可依本发明的揭示适当地将本发明应用于其它类型的调制技术。Those skilled in the art can understand the implementation details and changes of the method inventions in FIG. 15 and FIG. 16 through the disclosure of the device invention in FIG. 3 and FIG. 5 . Furthermore, although the foregoing embodiments take quadrature amplitude modulation as an example, this is not a limitation to the present invention, and those skilled in the art can appropriately apply the present invention to other types of modulation techniques according to the disclosure of the present invention.
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510032273.9ACN105871396B (en) | 2015-01-22 | 2015-01-22 | Signal receiving end capable of adaptively adjusting soft information and signal processing method thereof |
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510032273.9ACN105871396B (en) | 2015-01-22 | 2015-01-22 | Signal receiving end capable of adaptively adjusting soft information and signal processing method thereof |
| Publication Number | Publication Date |
|---|---|
| CN105871396A CN105871396A (en) | 2016-08-17 |
| CN105871396Btrue CN105871396B (en) | 2018-05-15 |
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510032273.9AExpired - Fee RelatedCN105871396B (en) | 2015-01-22 | 2015-01-22 | Signal receiving end capable of adaptively adjusting soft information and signal processing method thereof |
| Country | Link |
|---|---|
| CN (1) | CN105871396B (en) |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111858138B (en)* | 2020-07-08 | 2023-09-19 | 上海威固信息技术股份有限公司 | LDPC decoding optimization method based on three-dimensional TLC flash memory page unbalance bit errors |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1404656A (en)* | 2000-12-04 | 2003-03-19 | 皇家菲利浦电子有限公司 | Scaling of demodulated data in an interleaver memory |
| CN1627415A (en)* | 2003-11-28 | 2005-06-15 | 株式会社东芝 | Method and apparatus for data reproducing using iterative decoding in a disk drive |
| CN1839577A (en)* | 2003-07-29 | 2006-09-27 | 高通股份有限公司 | Scaling and Quantization of Soft Decision Metrics for Decoding |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2003075528A1 (en)* | 2002-03-07 | 2003-09-12 | Telefonaktiebolaget L M Ericsson (Publ) | Soft value calculation for multilevel signals |
| US9130589B2 (en)* | 2012-12-19 | 2015-09-08 | Avago Technologies General Ip (Singapore) Pte. Ltd. | Low density parity check decoder with dynamic scaling |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1404656A (en)* | 2000-12-04 | 2003-03-19 | 皇家菲利浦电子有限公司 | Scaling of demodulated data in an interleaver memory |
| CN1839577A (en)* | 2003-07-29 | 2006-09-27 | 高通股份有限公司 | Scaling and Quantization of Soft Decision Metrics for Decoding |
| CN1627415A (en)* | 2003-11-28 | 2005-06-15 | 株式会社东芝 | Method and apparatus for data reproducing using iterative decoding in a disk drive |
| Publication number | Publication date |
|---|---|
| CN105871396A (en) | 2016-08-17 |
| Publication | Publication Date | Title |
|---|---|---|
| EP1946509B1 (en) | Method and apparatus for normalizing input metric to a channel decoder in a wireless communication system | |
| TW200913596A (en) | Soft-bit de-mapping device and method of generating soft bits for decoding | |
| RU2328081C2 (en) | Hybrid correction method and unit with decision feedback | |
| JP2006515483A (en) | Soft information scaling for iterative decoding | |
| US7979032B2 (en) | Estimating statistical properties of noise in modulated data carrier signals | |
| EP1985026A2 (en) | Wireless communication unit and method for receiving a wireless signal | |
| JP4763057B2 (en) | Method and apparatus for normalizing metrics input to channel decoder in wireless communication system | |
| US20160191214A1 (en) | Method and apparatus for estimating channel information | |
| CN105871396B (en) | Signal receiving end capable of adaptively adjusting soft information and signal processing method thereof | |
| US7826523B2 (en) | Effective adaptive filtering techniques | |
| CN1917488A (en) | Bit Error Rate Estimation System and Related Methods | |
| TWI555361B (en) | Signal receiving end with adaptive soft information adjustment and its associated signal processing method | |
| CN108781129A (en) | Log-likelihood calculations circuit, reception device and log-likelihood calculations method | |
| US9979567B2 (en) | Equalization enhancing module, demodulation system and equalization enhancing method | |
| US8358683B2 (en) | Channel equalizer | |
| CN107204954B (en) | A kind of optimization method and system of fixed point Soft Inform ation | |
| JP5478327B2 (en) | Wireless communication system, receiver, and demodulation method for transmitting and receiving signals generated by modulo arithmetic | |
| TWI540858B (en) | Wireless receiving device and signal processing method thereof | |
| CN116318542B (en) | An adaptive modulation and coding method based on LDPC code | |
| RU2450451C2 (en) | Logic for quick calculation of square root for correction in mimo system | |
| CN101682608B (en) | Method of demapping data symbols for decoding in a wired communication system | |
| US8705662B2 (en) | Soft decision method and associated signal receiving system | |
| KR20090003636A (en) | Apparatus and method for symbol demapping in wireless communication system and receiver for same | |
| CN105991148B (en) | Wireless receiving device and signal processing method thereof | |
| CN107294888B (en) | Equalization enhancement module, demodulation system and equalization enhancement method |
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right | Effective date of registration:20211104 Address after:China Taiwan Hsinchu Science Park Hsinchu city Dusing Road No. 1 Patentee after:MEDIATEK Inc. Address before:Taiwan Hsinchu County Tai Yuan Street China jhubei City, No. 26 4 floor 1 Patentee before:MSTAR SEMICONDUCTOR Inc. | |
| TR01 | Transfer of patent right | ||
| CF01 | Termination of patent right due to non-payment of annual fee | Granted publication date:20180515 Termination date:20220122 | |
| CF01 | Termination of patent right due to non-payment of annual fee |