Orthogonal time-frequency two-dimensional space modulation signal receiving method and receiver
阅读说明：本技术 一种正交时频二维空间调制信号接收方法及接收器 (Orthogonal time-frequency two-dimensional space modulation signal receiving method and receiver ) 是由 牛凯 戴金晟 徐晋 于 2020-04-26 设计创作，主要内容包括：本发明公开了一种正交时频二维空间调制信号接收方法及接收器,通过对接收信号进行Zak变换解调,得到时延-多普勒域接收信号,对时延-多普勒域接收信号进行重排列得到加噪码字,采用基于因子图的期望传播算法,通过迭代方式计算各码字的译码概率,再根据各码字的译码概率进行判决,将译码概率最大的码字作为译码结果输出,由于在最终的检测算法上使用了期望传播算法,在迭代过程中大大减少了概率乘积与加和运算,使得整体系统的性能有所提升,在保障了译码的准确度的同时,降低了系统的运算复杂度,更加符合实际需要。(The invention discloses a method and a receiver for receiving orthogonal time-frequency two-dimensional space modulation signals, which are used for obtaining time delay-Doppler domain received signals by performing Zak transform demodulation on the received signals, rearranging the time delay-Doppler domain received signals to obtain noise-added code words, calculating the decoding probability of each code word in an iterative mode by adopting an expected propagation algorithm based on a factor graph, judging according to the decoding probability of each code word, and outputting the code word with the maximum decoding probability as a decoding result.)
1. A method for receiving orthogonal time-frequency two-dimensional space modulation signals is characterized by comprising the following steps:
performing Zak transform demodulation on the received signal to obtain a delay-Doppler domain received signal;
rearranging the time delay-Doppler domain received signals to obtain noise-added code words, wherein the noise-added code words are the results of the code words after noise is superimposed through a channel;
calculating the decoding probability of each code word in an iterative mode by adopting an expected propagation algorithm based on a factor graph;
and judging according to the decoding probability of each code word, and outputting the code word with the maximum decoding probability as a decoding result.
2. The method for receiving orthogonal time-frequency two-dimensional space modulation signals according to claim 1, wherein the performing Zak transform demodulation on the received signals to obtain the delay-doppler domain received signals comprises:
carrying out Viger transform on the received signal, and sampling to obtain a time-frequency domain two-dimensional sampling signal point;
and transforming the time-frequency domain two-dimensional sampling signal points to a time delay-Doppler domain to obtain a time delay-Doppler domain receiving signal.
3. The orthogonal time-frequency two-dimensional space modulation signal receiving method according to claim 1, wherein the code words comprise one-dimensional vectors of delay-doppler domain transmission signal sequences, one-dimensional vectors of delay-doppler domain reception signal sequences, and one-dimensional vectors of noise in delay-doppler domain reception signals, the factor graph comprises factor nodes and check nodes, the factor nodes represent the code words, the calculating the decoding probability of each code word in an iterative manner comprises,
a. initializing iteration parameters;
b. updating the mean and variance of each factor node;
c. respectively calculating the mean value and the variance transmitted from each factor node to the check node according to the mean value and the variance of each factor node;
d. updating the mean value and the variance of the check node after receiving the information according to the mean value and the variance transmitted to the check node by each factor node;
e. calculating the mean value and the variance of the check node returned to the factor node according to the updated mean value and the variance of the check node;
f. updating the mean value and the variance of the estimated factor node according to the mean value and the variance transmitted back to the factor node by the check node;
g. calculating the probability of each code word of the factor node according to the mean value and the variance of the estimated factor node;
h. and b, iterating the steps b-g until the iteration times reach a set value, jumping out of a loop, and outputting the decoding probability of each code word.
4. The method for receiving orthogonal time-frequency two-dimensional space modulation signals according to claim 3, wherein the initializing the iteration parameters comprises:
setting the probability of each code word of the initial factor node to be the same, initializing the return variance of the check node to be infinite, setting the return mean value to be 0, and setting the iteration counter to be 1.
5. The method for receiving orthogonal time-frequency two-dimensional space modulation signals according to claim 3, wherein the calculating of the mean and variance of the factor node to the check node according to the mean and variance of the factor node conforms to the following formula:
whereinRepresents the mean and variance of the ith factor node in the t iteration,
6. The method for receiving orthogonal time-frequency two-dimensional space modulation signals according to claim 3, wherein the calculating of the mean and variance of the check nodes returning to the factor nodes according to the mean and variance of the check nodes satisfies the following formula
Wherein the content of the first and second substances,
7. The method for receiving orthogonal time-frequency two-dimensional space modulation signals according to claim 3, wherein the probability of each code word of the calculation factor node is in accordance with the following formula according to the mean and variance of the estimation factor node
Wherein the content of the first and second substances,is the variance of the ith factor node,
8. An orthogonal time-frequency two-dimensional spatial modulation signal receiver, comprising:
the Zak demodulation module is used for performing Zak demodulation on the received signal to obtain a delay-Doppler domain received signal;
the rearrangement module is used for rearranging the time delay-Doppler domain received signals to obtain a noise-added code word;
the probability calculation module is used for calculating the decoding probability of each code word in an iterative mode by adopting an expected propagation algorithm based on a factor graph;
and the decoding judgment module is used for judging according to the decoding probability of each code word calculated by the probability calculation module and outputting the code word with the maximum decoding probability as a decoding result.
9. The orthogonal time-frequency two-dimensional spatial modulation signal receiver of claim 8, wherein the Zak demodulation module comprises:
the sampling module is used for carrying out the Wigner transformation on the received signals and sampling to obtain time-frequency domain two-dimensional sampling signal points;
and the octave Fourier transform module is used for performing octave Fourier transform on the time-frequency domain two-dimensional sampling signal points sampled by the sampling module to obtain a time delay-Doppler domain receiving signal.
Technical Field
The invention relates to the technical field of mobile communication, in particular to a method and a receiver for receiving orthogonal time-frequency two-dimensional space modulation signals.
Background
In wireless communication, a waveform design technique is often used to resist the influence of a complex wireless communication environment on communication quality and to improve the transmission rate and the bandwidth utilization rate of communication. Good waveform design is helpful to improve communication quality and communication speed, for example, Orthogonal Frequency Division Multiple Access (OFDMA) access technology used in 4G can well resist frequency selective fading in a wireless transmission environment, and can obtain high spectrum utilization rate. However, the OFDMA technology is based on Orthogonal Frequency Division Multiplexing (OFDM) modulation, and the sensitivity of the OFDMA technology to the doppler effect is extremely high, when the terminal moves at a high speed, the doppler effect will cause that each subcarrier allocated by the OFDM is no longer orthogonal, which generates intersymbol interference, and greatly affects the communication quality;
in view of this, the Orthogonal Time Frequency Spreading (OTFS) technology considers the characteristics of a wireless channel, and carries information in a delay-doppler domain for transmission, and for a mobile terminal, because the channel response in the delay-doppler domain caused by the movement of the mobile terminal is multiple impacts, the information is carried therein without generating intersymbol interference, and thus, more excellent communication performance can be obtained;
however, due to the transformation between the time delay-doppler domain and the time-frequency domain and the noise influence contained in the finally received signal, the traditional signal detection algorithm cannot meet the actual requirement, and if the maximum likelihood detection algorithm is used, the calculation complexity is high.
Disclosure of Invention
In view of this, the present invention provides a method for receiving an orthogonal time-frequency two-dimensional space modulation signal, including:
performing Zak transform demodulation on the received signal to obtain a delay-Doppler domain received signal;
rearranging the time delay-Doppler domain received signals to obtain noise-added code words, wherein the noise-added code words are the results of the code words after noise is superimposed through a channel;
calculating the decoding probability of each code word in an iterative mode by adopting an expected propagation algorithm based on a factor graph;
and judging according to the decoding probability of each code word, and outputting the code word with the maximum decoding probability as a decoding result.
Preferably, performing Zak transform demodulation on the received signal to obtain a delay-doppler domain received signal includes:
carrying out Viger transform on the received signal, and sampling to obtain a time-frequency domain two-dimensional sampling signal point;
and transforming the time-frequency domain two-dimensional sampling signal points to a time delay-Doppler domain to obtain a time delay-Doppler domain receiving signal.
Preferably, the noise-added code words include a one-dimensional vector of a delay-doppler domain transmission signal sequence, a one-dimensional vector of a delay-doppler domain reception signal sequence, and a one-dimensional vector of noise in a delay-doppler domain reception signal, the factor graph includes factor nodes and check nodes, the factor nodes represent the noise-added code words, the calculating the decoding probability of each code word in an iterative manner includes,
a. initializing iteration parameters;
b. updating the mean and variance of each factor node;
c. respectively calculating the mean value and the variance transmitted from each factor node to the check node according to the mean value and the variance of each factor node;
d. updating the mean value and the variance of the check node after receiving the information according to the mean value and the variance transmitted to the check node by each factor node;
e. calculating the mean value and the variance of the check node returned to the factor node according to the updated mean value and the variance of the check node;
f. updating the mean value and the variance of the estimated factor node according to the mean value and the variance transmitted back to the factor node by the check node;
g. calculating the probability of each code word of the factor node according to the mean value and the variance of the estimated factor node;
h. and b, iterating the steps b-g until the iteration times reach a set value, jumping out of a loop, and outputting the decoding probability of each code word.
Preferably, initializing the iteration parameters comprises:
setting the probability of each code word of the initial factor node to be the same, initializing the return variance of the check node to be infinite, setting the return mean value to be 0, and setting the iteration counter to be 1.
Preferably, according to the mean and variance of the factor node, the mean and variance of the transmission from the factor node to the check node are calculated according to the following formula:
whereinRepresents the mean and variance of the ith factor node in the t iteration,denotes the mean and variance, h, passed from the ith factor node to the jth check nodej,iRepresenting the parameters of the jth row and ith column of the channel matrix H.
Preferably, according to the mean and variance of the check node, calculating the mean and variance of the check node returning to the factor node according to the following formula
Wherein the content of the first and second substances,respectively table j shows the variance and mean of the check nodes,
Preferably, the probability of calculating each code word of the factor node according to the mean and variance of the estimated factor node is in accordance with the following formula
Wherein the content of the first and second substances,is the variance of the ith factor node,is the average of the ith factor node,
An orthogonal time-frequency two-dimensional spatial modulation signal receiver, comprising:
the Zak demodulation module is used for performing Zak demodulation on the received signal to obtain a delay-Doppler domain received signal;
the rearrangement module is used for rearranging the time delay-Doppler domain received signals to obtain a noise-added code word;
the probability calculation module is used for calculating the decoding probability of each code word in an iterative mode by adopting an expected propagation algorithm based on a factor graph;
and the decoding judgment module is used for judging according to the decoding probability of each code word calculated by the probability calculation module and outputting the code word with the maximum decoding probability as a decoding result.
Preferably, the Zak demodulation module comprises:
the sampling module is used for carrying out the Wigner transformation on the received signals and sampling to obtain time-frequency domain two-dimensional sampling signal points;
and the octave Fourier transform module is used for performing octave Fourier transform on the time-frequency domain two-dimensional sampling signal points sampled by the sampling module to obtain a time delay-Doppler domain receiving signal.
From the above, it can be seen that, according to the orthogonal time-frequency two-dimensional space modulation signal receiving method and the receiver provided by the present invention, Zak transform demodulation is performed on the received signal to obtain the delay-doppler domain received signal, rearrangement is performed on the delay-doppler domain received signal to obtain the noisy codeword, an expected propagation algorithm based on a factor graph is adopted, the decoding probability of each codeword is calculated in an iterative manner, then a decision is performed according to the decoding probability of each codeword, and the codeword with the largest decoding probability is output as a decoding result.
Drawings
FIG. 1 is a schematic diagram of an orthogonal time-frequency spreading communication system according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of an orthogonal time-frequency spreading communication flow according to an embodiment of the present invention;
FIG. 3 is a block diagram of an EP algorithm flow according to an embodiment of the present invention, wherein node update refers to updating the mean and variance of corresponding nodes;
fig. 4 is a schematic diagram illustrating a case where a factor node transmits a message to all check nodes when EP algorithm detection is performed according to the embodiment of the present invention;
fig. 5 is a schematic diagram illustrating a case where a check node transmits a message to all factor nodes when performing EP algorithm detection according to the embodiment of the present invention;
FIG. 6 is a block diagram of a flow chart of a method for receiving an orthogonal time-frequency two-dimensional space modulation signal according to an embodiment of the present invention;
fig. 7 is a schematic diagram of an orthogonal time-frequency two-dimensional space modulation signal receiver module according to an embodiment of the invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to specific embodiments and the accompanying drawings.
It should be noted that all expressions using "first" and "second" in the embodiments of the present invention are used for distinguishing two entities with the same name but different names or different parameters, and it should be noted that "first" and "second" are merely for convenience of description and should not be construed as limitations of the embodiments of the present invention, and they are not described in any more detail in the following embodiments.
As described above, the orthogonal time-frequency spreading technology (OTFS) has a relatively high capability of resisting doppler, and can obtain a more excellent performance, and as shown in fig. 1, the orthogonal time-frequency spreading communication system includes a transmitting end 100 and a receiving end 102, where the transmitting end 100 and the receiving end 102 transmit signals through a channel 101, and the receiving end 102 includes a Zak demodulation module 1021, a probability calculation module 1022, and a decoding decision module 1023.
As shown in fig. 2, the orthogonal time-frequency spreading communication process is that the transmitting end 100 performs inverse octave fourier transform (ISSFT) on the two-dimensional signal in the delay-doppler domain, so that the two-dimensional signal in the delay-doppler domain is transformed into a time-frequency domain, and a two-dimensional signal in the time-frequency domain X [ n, m ] is obtained, where the ISSFT formula is as follows:
wherein, the number of sub-carriers and the number of OTFS symbols are M and N respectively, and the total number of transmission symbols K is NM
And then carrying out Hisenberg transformation on the obtained X [ n, m ] to transform the X [ n, m ] into a time domain, wherein a transformation result s (t) is a physically-realizable time domain one-dimensional signal, and a Hisenberg transformation formula is as follows:
where T is an Orthogonal Frequency Division Multiplexing (OFDM) symbol length, Δ f is a subcarrier spacing, and both satisfy T Δ f ═ 1.
At this point, the transmitting end 100 completes the modulation of the signal.
The transform result s (t) modulated by the transmitting end 100 passes through the channel 101, according to the assumed time-varying channel 101 model, H (t, f) can be obtained by H (τ, ν) through the ISSFT transform, and r (t) is a time-domain signal (without considering noise influence) after passing through the channel 101, and the formula passing through the channel 101 is as follows:
r(t)＝∫H(t,f)S(f)ej2πftdf
after receiving end 102 receives signal r (t), sampling module in Zak demodulation module 1021 performs wigner transformation on received signal r (t), the process of wigner transformation is equivalent to passing through a matched filter, the obtained time-frequency domain two-dimensional sampling signal point Y [ n, m ] corresponds to X [ n, m ] of sending end 100, and the formula of the wigner transformation is as follows:
in digital domain, the signal is sampled to obtain time-frequency domain two-dimensional sampling signal points
Y[n,m]＝Y(t,f)|t＝nT,f＝mΔf
It should be noted that, in particular, if the transmission filter g is usedtxAnd a receiving filter grxThe method is completely localized in time and frequency, satisfies the biorthogonal characteristic of time-frequency domain, and can be simplified into Y [ n, m ] by the original formula]＝H[n,m]X[n,m]
Where H [ n, m ] is the corresponding sampling value of the time-frequency domain impulse of the channel 101
H[n,m]＝∫∫h(τ,ν)ej2πνnTe-j2πmΔfτdτdν
Then, a symplectic fourier transform module in the Zak demodulation module 1021 transforms the sampled signal to a delay-doppler domain to obtain a delay-doppler domain received signal for signal detection, where the symplectic fourier transform (SSFT) formula is as follows:
the delay-Doppler domain received signal is actually the result of two-dimensional cyclic convolution between the delay-Doppler domain transmitted signal and the impulse response of the delay-Doppler domain channel 101, i.e. the delay-Doppler domain received signal is
For the delay-doppler domain received signal obtained after demodulation, it is actually the result of two-dimensional cyclic convolution between the delay-doppler domain transmitted signal and the impulse response of the delay-doppler domain channel 101 and superposition of noise, and the specific form is
The detection of the information symbol sequence x [ k, l ] at the delay-doppler domain transmitter 100 requires a rearrangement operation to obtain a noisy codeword. I.e. x k, l is the solution of the following set of MN linear equations
y＝Hx+w
Wherein, y, x, w are the one-dimensional vector representation of y [ k, l ], x [ k, l ], w [ k, l ], H is the channel 101 parameter matrix, i.e. y is the code word of adding noise, x is the code word, the code word of adding noise is the result of the code word after the noise is superimposed through the channel.
For example, the noise code word includes a one-dimensional vector x [ k, l ] of the delay-doppler domain transmission signal sequence, a one-dimensional vector y [ k, l ] of the delay-doppler domain reception signal sequence, and a one-dimensional vector w [ k, l ] of noise in the delay-doppler domain reception signal;
representing the code words by factor nodes in a factor graph;
the EP algorithm is shown in fig. 3 and includes:
s301, acquiring a received signal and a channel parameter;
s302, the iteration parameters are initialized, including the assumption that the probability of each code word taken by the initial factor node is the same, namelyWherein Q is a modulation order number, and Q is a modulation order number,code word c representing the ith factor nodeiThe probability of (c). Initializing check node pass-back variance
s303, initializing an iteration counter, and setting an iteration count L to be 1;
s304 updates the factor nodes and passes parameters to the check nodes, including:
mean and variance of update factor nodes
Wherein
Calculating the mean value and the variance transmitted from the factor node to the check node according to the mean value and the variance of the factor node;
wherein
the process is repeated until the calculation of all factor nodes transmitting to the check node is completed, if the first process is executed, the factor node serial number i is set to 1, and the factor node serial number i is updated to i +1 once each time the process is completed, and the calculation is completed until i is greater than MN.
S305 updates the check node and returns the parameters to the factor node, including:
according to the mean value and the variance transmitted from the factor node to the check node, the check node updates the mean value and the variance of the check node after receiving the information
Wherein
Calculating the mean value and the variance of the check node returned to the factor node according to the mean value and the variance of the check node
Wherein The process of the check node transmitting the mean and variance back to the factor node is shown in fig. 5;
repeating the process until the calculation of the mean value and the variance returned from all the check nodes to the factor node is completed, if the first process is executed, setting the serial number j of the factor node to be 1, updating the serial number j of the factor node to be j +1 once each time the process is completed, and completing the calculation until j is greater than MN.
S306 updating the estimation factor node and calculating the probability of each codeword, including:
updating the mean value and the variance of the estimated factor node according to the mean value and the variance returned by the check node to the factor node
Wherein the content of the first and second substances,is the variance of the ith factor node,is the average of the ith factor node.
Calculating the probability of each code word of the factor node according to the mean value and the variance of the estimated factor node
For calculating a function for the probability, in particular
Repeating the process until the calculation of the probability of each code word of all the factor nodes is completed, if the first process is executed, setting the serial number i of the factor node to be 1, updating the serial number i of the factor node to be i +1 once each time the first process is completed, and completing the calculation until i is larger than MN.
S307, after each iteration is finished, counting the number of iterations by L + 1;
s308, judging whether the iteration times reach a preset value, if not, repeating the iteration detection, if so, ending the iteration detection, and outputting the decoding probability.
Then according to the decoding probability of each factor node
The time delay-Doppler domain channel response has quasi-static and sparse characteristics. Data placement in the delay-doppler domain bearer does not suffer from the severe performance degradation that occurs when placed in the conventional time-frequency domain bearer. Assuming that the communication device at the mobile end is moving, when the velocity changes little, the doppler shift generated by the movement is stable near a certain value, and in addition, because the environment of the communication device changes little under the macro scale (compared with the distance between the communication device and the base station), the communication delay is also stable at a certain value, which is reflected on the channel matrix, that is, the delay-doppler domain channel matrix is a quasi-static sparse matrix, which is more stable than the dynamic channel matrix of the time-frequency domain when transmitting signals, thereby avoiding the performance deterioration caused by the dynamic change of the channel matrix.
Compared with the prior art, the embodiment uses an Expected Propagation (EP) algorithm on the final detection algorithm. The EP algorithm greatly reduces probability product and probability summation operation in an iteration process by assuming factor nodes and check nodes as Gaussian variables, reduces operation complexity, and simultaneously selects a better approximation formula through testing when approximate processing is performed on the transmitted Gaussian variables, thereby ensuring the accuracy of decoding.
In practical applications, a matrix or a vector is mostly used as a unit for data processing, and the embodiment is further described in detail in the form of a matrix.
The sending end processes:
and performing ISSFT transformation, Heisenberg transformation and pulse forming at a transmitting end to obtain an S matrix.
WhereinFor the ISSFT transform matrix (X is the initial signal matrix),
Expressed in vector form as:
channel processing:
according to the channel model, the variation of the transmitted signal at the channel is as follows (vector form):
r＝Hs+w
wherein
hiFor the channel parameters, Π is the forward cyclic permutation matrix and Δ is the diagonal matrix. The concrete structure is as follows:
processing at a receiving end: after receiving the signal, the inverse process of the transmitting end is executed on the signal, and the formula is as follows:
wherein the content of the first and second substances,for SFFT transformation matrix, FMFor a Virgener transformation matrix, GrxIs a pulse-shaped matrix.
Expressed in vector form as
Bringing r into the simplification of y
If a rectangular window is used, there is Gtx＝Grx＝IMThe above formula becomes
Wherein, P(i)And Q(i)The effective channel parameter matrixes respectively represent the rectangular pulse time delay and the Doppler domain.
And (3) detection and judgment:
the relation between the output signal vector and the input signal vector is obtained in the sending end
For convenience of description, in this step abstract is
y＝Hx+w。
The EP detection algorithm flow is as follows:
parameter initialization
Where j is 0,1.. MN, i is 0,1.. MN denotes a check node and a factor node, respectively. L is the set cycle number, and the factor node is xiIndicating that check nodes are denoted by fjAnd (4) showing. Mean and variance of factor nodes are represented byAndthe mean and variance of the check nodes are denoted by z and v the mean and variance of the final estimation factor nodes are denoted by ξ and γ.
Updating factor nodes and transmitting messages to check nodes, the transmitted messages including two parameters of mean value and varianceAnd
And calculating the mean and the variance of the ith factor node transferred to the jth check node.
WhereinA complex transpose of the parameters representing the ith column and jth row of the channel matrix H.
This process is repeated until all computations for passing messages between nodes are completed.
Updating check nodes and transmitting messages to factor nodes, the transmitted messages including two parameters of mean and variance, i.e.Andcalculating the mean value and variance of each check node
And calculating the mean and the variance of the j check node to the i factor node.
This process is repeated until all computations for passing messages between nodes are completed.
Estimating the mean and variance of the final variable based on the mean and variance of the check node transmissions and calculating therefrom the probability of the original signal
The final mean and variance of the variables are calculated as follows
And calculating the probability of each code word of each factor node according to the variance and the mean of the estimated factor nodes.
Wherein
And detecting whether the iteration times are reached, if not, repeating the iteration, and if so, outputting the decoding probability.
And then, judging according to the decoding probability of each factor node, and outputting a code word with the maximum decoding result.
Based on the above research results, an embodiment of the present invention provides a method for receiving an orthogonal time-frequency two-dimensional space modulation signal, as shown in fig. 6, where the specific implementation process of the method includes:
s701, performing Zak transform demodulation on the received signal to obtain a delay-Doppler domain received signal;
wherein, the Zak transform demodulation to obtain the delay-doppler domain received signal comprises:
carrying out Viger transform on the received signal, and sampling to obtain a time-frequency domain two-dimensional sampling signal point;
and transforming the time-frequency domain two-dimensional sampling signal points to a time delay-Doppler domain to obtain a time delay-Doppler domain receiving signal.
S702, rearranging the time delay-Doppler domain received signals to obtain noise-added code words;
for example, the noise-added code words include a one-dimensional vector of the delay-doppler domain transmission signal sequence, a one-dimensional vector of the delay-doppler domain reception signal sequence, and a one-dimensional vector of noise in the delay-doppler domain reception signal.
S703, calculating the decoding probability of each code word in an iterative mode by adopting an expected propagation algorithm based on a factor graph;
s704, the judgment is carried out according to the decoding probability of each code word, and the code word with the maximum decoding probability is output as a decoding result.
According to the invention, the Zak transform demodulation is carried out on the received signals to obtain the time delay-Doppler domain received signals, the time delay-Doppler domain received signals are rearranged to obtain the noise-added code words, an expected propagation algorithm based on a factor graph is adopted, the decoding probability of each code word is calculated in an iterative mode, then the judgment is carried out according to the decoding probability of each code word, and the code word with the maximum decoding probability is output as the decoding result.
In the embodiment of the invention, the factor graph comprises factor nodes and check nodes, wherein the factor nodes represent code words, the decoding probability of each code word is calculated in an iterative mode,
a. initializing iteration parameters;
b. updating the mean and variance of each factor node;
c. respectively calculating the mean value and the variance transmitted from each factor node to the check node according to the mean value and the variance of each factor node;
d. updating the mean value and the variance of the check node after receiving the information according to the mean value and the variance transmitted to the check node by each factor node;
e. calculating the mean value and the variance of the check node returned to the factor node according to the updated mean value and the variance of the check node;
f. updating the mean value and the variance of the estimated factor node according to the mean value and the variance transmitted back to the factor node by the check node;
g. calculating the probability of each code word of the factor node according to the mean value and the variance of the estimated factor node;
h. and b, iterating the steps b-g until the iteration times reach a set value, jumping out of a loop, and outputting the decoding probability of each code word.
In an embodiment of the present invention, initializing the iteration parameters comprises:
setting the probability of each code word of the initial factor node to be the same, initializing the return variance of the check node to be infinite, setting the return mean value to be 0, and setting the iteration counter to be 1.
In the embodiment of the invention, according to the mean value and the variance of the factor node, the mean value and the variance transmitted from the factor node to the check node are calculated to conform to the following formula:
wherein
In the embodiment of the invention, according to the mean value and the variance of the check node, the mean value and the variance of the check node returning to the factor node are calculated to accord with the following formula
Wherein the content of the first and second substances,respectively table j shows the variance and mean of the check nodes,represents the mean and variance, h, passed from the jth check node to the ith factor nodej,iRepresenting the parameters of the jth row and ith column of the channel matrix H.
In the embodiment of the invention, according to the mean value and the variance of the estimation factor node, the probability of each code word of the calculation factor node accords with the following formula
Wherein the content of the first and second substances,
An embodiment of the present invention further provides an orthogonal time-frequency two-dimensional space modulation signal receiver, as shown in fig. 7, including:
a Zak demodulation module 801, configured to perform Zak demodulation on the received signal to obtain a delay-doppler domain received signal;
wherein, Zak demodulation module 801 includes:
a sampling module 8011, configured to perform wigner transformation on a received signal, and sample to obtain a time-frequency domain two-dimensional sampling signal point;
a fourier transform module 8012, configured to perform fourier transform on the time-frequency domain two-dimensional sampling signal points sampled by the sampling module 8011, so as to obtain a delay-doppler domain received signal.
And a rearrangement module 802, configured to rearrange the delay-doppler domain received signals to obtain a noisy codeword.
And a probability calculation module 803, configured to calculate, in an iterative manner, the decoding probability of each codeword by using a desired propagation algorithm based on a factor graph.
A decoding decision module 804, configured to decide according to the decoding probability of each codeword calculated by the probability calculation module 802, and output the codeword with the largest decoding probability as a decoding result.
Those of ordinary skill in the art will understand that: the discussion of any embodiment above is meant to be exemplary only, and is not intended to intimate that the scope of the disclosure, including the claims, is limited to these examples; within the idea of the invention, also features in the above embodiments or in different embodiments may be combined, steps may be implemented in any order, and there are many other variations of the different aspects of the invention as described above, which are not provided in detail for the sake of brevity.
While the present invention has been described in conjunction with specific embodiments thereof, many alternatives, modifications, and variations of these embodiments will be apparent to those of ordinary skill in the art in light of the foregoing description. For example, other memory architectures (e.g., dynamic ram (dram)) may use the discussed embodiments.
The embodiments of the invention are intended to embrace all such alternatives, modifications and variances that fall within the broad scope of the appended claims. Therefore, any omissions, modifications, substitutions, improvements and the like that may be made without departing from the spirit and principles of the invention are intended to be included within the scope of the invention.