A bit-level communication probability shaping method and system for NOMA

Abstract

The present invention relates to the field of optical fiber communication technology, and discloses a bit-level communication probability shaping method and system for NOMA, wherein the shaping method includes the following steps: two different users transmit data and modulate them into QPSK-OFDM signals; for one of the QPSK-OFDM signals, implement symbol probability transformation through improved BWDM; perform power allocation and superposition for the two QPSK-OFDM signals to form a NOMA-OFDM signal; after receiving the signal, perform channel estimation and equalization on the signal; perform regional judgment on the equalized signal, and restore the signal respectively through the regional judgment method of high-power and low-power signals; and obtain the restored low-power QPSK-OFDM signal by deconstructing BWDM. The present invention is combined with an improved bit-level probability shaping method, which avoids the lengthy calculation and error propagation problems of the SIC technology, and can also obtain more accurate signal recovery, and further simplifies the signal judgment process, reducing the coding complexity of the system.

Description

Technical Field

[0001] The present invention relates to the field of optical fiber communication technology, and in particular to a NOMA-oriented bit-level communication probability shaping method and system. Background Art

[0002] The sixth generation of mobile networks (6G) is booming and has become a hot area of ​​global research and development. As a catalyst for the advancement of communication technology, 6G is expected to achieve seamless global coverage. Compared with 5G, 6G's data transmission rate can be increased by fifty times, and the delay is reduced to one tenth. It is far superior to 5G in terms of peak rate, delay, traffic density and spectrum efficiency. As the number of users and smart devices continues to increase, the limited spectrum bandwidth needs to serve more terminal devices, which will lead to a serious decline in the service quality of each terminal.

[0003] In order to avoid this problem, non-orthogonal multiple access technology (NOMA) was proposed to improve the performance of the next generation of communication systems. As a key technology for optical fiber communication, it effectively solves the problems of resource utilization and limited number of users brought by orthogonal multiple access technology (OMA). Compared with high-capacity OMA technology, NOMA has many advantages such as higher user throughput, looser channel feedback, lower transmission delay and more flexible transmission strategy. Therefore, NOMA technology can enable the communication system to reach a better working point and optimize both spectrum efficiency and energy efficiency. These outstanding advantages just meet the demand for higher speed and lower latency communications in the upcoming 6G era, promote the popularization of the Internet of Everything, and therefore have attracted the attention of many researchers.

[0004] NOMA technology is usually implemented in the power domain, using power division multiplexing (PDM) to achieve non-orthogonal multiple access. This NOMA scheme can be integrated with other advanced orthogonal technologies, such as orthogonal frequency division multiplexing (NOMA-OFDM). Compared with traditional OFDM technology, this combination allows more user data to be transmitted in the power domain. In NOMA-OFDM, different OFDM signals are distinguished based on the allocated power. Generally speaking, the transmitter allocates different powers to each signal, and the receiver uses successive interference cancellation (SIC) to decode the signals in descending order of power. However, SIC still has some challenges: (1) Evaluating the number of users increases the complexity of the system; (2) Error propagation may occur, that is, a decoding error of one user will cause decoding errors for all subsequent users in the SIC decoding queue. Therefore, how to more efficiently recover the multiplexed signal and improve the transmission efficiency of the system is one of the problems that need to be solved.

[0005] As the optical signal-to-noise ratio (OSNR) continues to grow, the inherent power limitation problem in optical communication systems has led to an increase in the nonlinear effects of optical fibers, resulting in a significant decrease in transmission performance. Probabilistic shaping (PS) can enhance the system's adaptability to noise and other interference factors. PS aims to increase the distribution probability of internal constellation points while reducing the distribution probability of external constellation points, significantly reducing the average constellation power. At present, some research results around the PS scheme have been widely used, such as component distribution matching, layered distribution matching, and prefix-free code distribution matching. Whether in short-distance intensity modulation / direct detection (IM / DD) communication systems or long-distance communication systems, the combination of PS and other coding technologies has significantly improved the transmission performance of the system. In the NOMA system, the labels representing the expected probabilities in schemes such as symbol-level distribution matching still need to be converted or mapped to binary sequences to be compatible with the forward error correction (FEC) coding and high-order modulation process of the cascaded block error correction code (LDPC). The encoding and decoding process of this symbol-level traditional PS scheme introduces high computational complexity to the system, increases system power consumption, and increases the difficulty of hardware implementation. Therefore, in the NOMA system, which is very sensitive to resource allocation, how to reduce hardware resource consumption is an important problem.

[0006] At present, technologies such as lookup table, Sell mapping and hash index have been proposed to reduce the complexity of the system. Inspired by these hardware-efficient solutions, a low-complexity bit-level PS scheme was proposed to reduce the hardware resource requirements in the PS process. The bit-level scheme operates on bit data and is not constrained by the data structure throughout the process, allowing multiple bits to be flexibly combined. Specifically, its final output is a set of bits from the input data source, which can be seamlessly compatible with subsequent bit-level processes. Among them, 16QAM bit-weighted distribution matching (BWDM) is a typical bit-level PS technology. By changing the probability distribution of bit combinations, it avoids the lengthy probability calculation and allocation process of traditional probability shaping schemes.

[0007] There are also some NOMA-oriented fiber optic communication probability shaping methods and / or systems in the prior art, such as the Chinese invention patent with application number CN202011164050.5 and the name of the method for improving the transmission capacity of non-orthogonal multiple access visible light communication based on probability shaping, which can improve the transmission rate and user fairness by adjusting the probability distribution of the transmitted signal on the basis of non-orthogonal multiple access technology. The method proposed in this invention can not only narrow this gap, but also reduce the achievable rate gap between users (i.e. improve fairness). However, this invention uses a method of mapping binary data into Maxwell-Boltzmann distribution symbols to achieve constellation probability shaping, and uses SIC for signal recovery. It is based on a visible light communication system, which cannot reduce the complexity of the system algorithm and the difficulty of hardware implementation, and cannot effectively avoid the problem of SIC error propagation. Summary of the invention

[0008] In order to overcome the shortcomings of the prior art, the purpose of the present invention is to provide a bit-level communication probability shaping method and system for NOMA, combined with an improved bit-level probability shaping method, by adopting BWDM to realize probability shaping at the bit level, that is, avoiding the lengthy calculation and error propagation problems of SIC technology, and can also obtain more accurate signal recovery, and further simplify the signal decision process, reducing the coding complexity of the system.

[0009] The present invention is implemented by the following technical solution: a bit-level communication probability shaping method for NOMA, comprising the following steps:

[0010] Step 1: The data transmitted by two different users are modulated into QPSK-OFDM signals;

[0011] Step 2: For one of the QPSK-OFDM signals, implement symbol probability transformation through improved BWDM;

[0012] Step 3: Power allocation and superposition of two QPSK-OFDM signals to form a NOMA-OFDM signal, where the signal in step 2 is allocated with low power;

[0013] Step 4: Transmit the NOMA-OFDM signal through optical fiber;

[0014] Step 5: After receiving the signal, perform channel estimation and equalization on the signal;

[0015] Step 6: Perform regional decision on the equalized signal, and restore the high-power QPSK-OFDM signal and the low-power QPSK-OFDM signal after symbol probability transformation respectively through the regional decision method of high-power and low-power signals;

[0016] Step 7: Get the recovered low-power QPSK-OFDM signal by de-BWDM.

[0017] Compared with the prior art, the present invention has the following beneficial effects:

[0018] 1. The present invention discloses a bit-level communication probability shaping method for NOMA, which adopts improved BWDM to realize probability shaping. Different from the traditional PS-16QAM scheme that needs to operate on 16 constellation points, the proposed method only needs to encode a small number of bits when processing low-power QPSK signals to realize the overall probability shaping of the superimposed 16QAM signal. It uses NOMA to increase the transmission capacity of the system, and realizes the probability shaping of the power multiplexing signal of the 16QAM modulation format by only changing the symbol distribution probability of the low-power signal. Through this method, the probability shaping process of the 16QAM signal is simplified and the coding complexity of the system is reduced.

[0019] 2. The present invention provides a bit-level communication probability shaping method for NOMA, which utilizes the traditional bit-level probability shaping principle and combines an improved bit-level probability shaping method on a low-power QPSK signal, and implements probability shaping at the bit level by adopting BWDM.

[0020] 3. The bit-level communication probability shaping method for NOMA of the present invention adopts a regional decision method, which avoids the lengthy calculation and error propagation problems of the SIC technology, can obtain more accurate signal recovery, and further simplifies the signal decision process.

[0021] 4. The bit-level communication probability shaping system for NOMA of the present invention has a simple module composition, low system hardware complexity, and does not use any block random access memory. By cooperating with the bit-level communication probability shaping method for NOMA, it can verify bit-level fiber-optic communication. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 It is a flow chart of the bit-level communication probability shaping method for NOMA of the present invention;

[0023] Figure 2 It is a flow chart of the improved BWDM in the present invention to realize symbol probability shaping;

[0024] Figure 3 It is a diagram showing the symbol probability distribution change of QPSK before and after bit flipping in the present invention;

[0025] Figure 4 It is a flow chart of NOMA-OFDM signal generation in the present invention;

[0026] Figure 5 It is a schematic diagram of superposition of OFDM signals of different powers in the present invention;

[0027] Figure 6 It is a schematic diagram of the regional decision principle of large and small power signals in the present invention;

[0028] Figure 7 It is a connection diagram of the bit-level communication probability shaping system for NOMA in the present invention;

[0029] Figure 8 It is the change diagram of generalized mutual information (GMI) of AWGN channel under different signal-to-noise ratio (SNR);

[0030] Fig. 9 This is a comparison chart of the bit error rate results of the multiplexed signal of the present invention and uniform 16QAM without probability shaping;

[0031] Fig.10It is a constellation diagram of the present invention and a bit error rate variation curve diagram of the signal under different k values.

[0032] In the figure: transmitter-1; receiver-2; coupler-3; optical fiber access device-4; multi-core optical fiber-5; optical fiber output device-6; waveform generator-11; electrical amplifier-12; modulator-13; optical fiber amplifier-14; laser-15; adjustable optical attenuator-21; photodiode-22; mixed signal oscilloscope-23. DETAILED DESCRIPTION

[0033] The following describes the embodiments of the present invention by specific examples, and those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the following embodiments and features in the embodiments can be combined with each other without conflict.

[0034] The purpose of the present invention is to provide an augmented reality glasses and an image processing method thereof in view of the defects of the prior art.

[0035] Embodiment 1

[0036] A bit-level communication probability shaping method for NOMA, referring to Figure 1 As shown, the following steps are included:

[0037] Step 1: Modulate the transmission data of two different users into QPSK-OFDM signals;

[0038] Step 2: For one of the QPSK-OFDM signals, implement symbol probability transformation through improved BWDM;

[0039] Step 3: Power allocation and superposition of two QPSK-OFDM signals to form a NOMA-OFDM signal, where the signal in step 2 is allocated with low power;

[0040] Step 4: Transmit the NOMA-OFDM signal through optical fiber;

[0041] Step 5: After receiving the signal, perform channel estimation and equalization on the signal;

[0042] Step 6: Perform regional decision on the equalized signal, and restore the high-power QPSK-OFDM signal and the low-power QPSK-OFDM signal after symbol probability transformation respectively through the regional decision method of high-power and low-power signals;

[0043] Step 7: Get the recovered low-power QPSK-OFDM signal by de-BWDM.

[0044] At the transmitter, the two OFDM signals with different powers in step 1 are linearly combined by superposition coding. Figure 5 As shown, both OFDM signals are QPSK modulated, and the superimposed signal is further mapped to 16QAM. On the basis of retaining the mapping rule of the second quadrant, the low-power signals of the first, fourth and third quadrants are superimposed in sequence by rotating 90°, 180° and 270° clockwise. The signal after power division multiplexing can be mathematically expressed as:

[0045]

[0046] In the formula, And s i and p i represents the signal and signal power of the i-th user.

[0047] In order to improve the quality of the transmission signal, the improved BWDM is used in the PDM-SDM system to realize the bit-level constellation probability shaping. BWDM is already applicable to the probability shaping of 16QAM. The original bit stream is first converted into four parallel bit sequences (G1-G4). Subsequently, G2 and G4 are divided into a group of k-bit bit combinations, and weighted and flipped according to the proportion of k bits. Finally, the probability of "1" or "0" appearing in G2 and G4 is increased (determined according to different mapping rules), and the probability of mapping constellation points in the inner circle of the constellation diagram will increase significantly. Based on the BWDM of 16QAM, the present invention improves this principle and applies it to the constellation diagram of the QPSK debugging format to realize the bit-level QPSK symbol distribution probability change.

[0048] At the receiving end, the low-power QPSK-OFDM signal after bit flipping is restored by using regional decision. Then, the signal is de-BWDMed. Specifically, the MSB placed after the sequence G2 data stream as a label is first extracted according to the BWDM principle of the transmitting end; then, the bit flip recovery is performed on each group of bits according to the MSB.

[0049] Embodiment 2

[0050] A bit-level communication probability shaping method for NOMA, specifically, referring to Figure 2 As shown, when k=2, the specific process of implementing symbol probability transformation of the QPSK-OFDM signal in step 2 through the improved BWDM includes:

[0051] Step 21: Divide the original bits of the M-bit serial low-power QPSK-OFDM signal into two parallel bit sequences (G1 and G2) of length M / 2 through cross mapping;

[0052] Step 22: Divide G2 into a group of k-bit bit combinations and weight them according to their original elements; add an additional bit as an additional significant weighted bit (MSB) in each group of bits as a decision weight; in the bit weighted decision process, if the number of "0"s in the k-bit is less than k / 2, the corresponding MSB is defined as "1".

[0053] Step 23: Flip the remaining bits in the bit combination bit by bit, otherwise, set the MSB to "0" and keep the original bit unchanged. The bit-by-bit inversion operation is a key step in BWDM and is used to increase the probability of the "0" bit in the sequence G2 before QPSK mapping.

[0054] Step 24: After bit flipping, the MSB is integrated and placed as a label in the sequence G2 data stream. After G1 and G2 perform parallel to serial (P / S) conversion, the data stream is modulated into QPSK-OFDM format.

[0055] By superimposing the two signals, the constellation diagram of the transmitted signal can be regarded as a 16QAM modulation format signal. Therefore, by changing the symbol distribution probability of the low-power QPSK signal, the probability of mapping to the constellation point in the inner circle of the 16QAM constellation diagram will increase significantly.

[0056] Reference Figure 3 As shown in the figure, the probability distribution of ("00", "10", "01", "11") changes from (0.25, 0.25, 0.25, 0.25) to (0.3738, 0.2539, 0.2527, 0.1222). Compared with the uniform 16QAM signal, the information entropy of the 16QAM signal after probability shaping is reduced from 4 to 3.9 bits / symbol. In addition, the proposed framework can also be combined with other probability shaping schemes, which increases the flexibility and versatility of the scheme.

[0057] In this embodiment, NOMA technology is used to implement s 1 and 2 The superposition of s 1 Defined as a high power signal, s 2 Defined as a low power signal, set the power allocation ratio p 1 / p 2 =4.

[0058] Embodiment 3

[0059] A bit-level communication probability shaping method for NOMA, specifically, referring to Figure 4 As shown, the NOMA-OFDM signal production process in step 3 includes:

[0060] Step 31: For two signals with different powers, the initial binary bit data is mapped to multiple subcarriers of the QPSK signal through serial-to-parallel conversion (S / P) to obtain an OFDM signal in the frequency domain;

[0061] Step 32: directly perform an inverse fast Fourier transform (IFFT) on one of the low-power OFDM signals to convert it from the frequency domain to the time domain; first insert a training sequence (TS) into the high-power OFDM signal for channel estimation and equalization at the receiving end, and then perform an inverse fast Fourier transform (IFFT) to convert it from the frequency domain to the time domain;

[0062] Step 33: Add a cyclic prefix (CP) in the time domain of the two OFDM signals to avoid inter-symbol interference;

[0063] Step 34: Perform parallel-to-serial conversion on the two signals respectively, perform power allocation, and obtain the NOMA-OFDM signal through superposition coding.

[0064] The signal channel estimation and equalization process in step 5 includes,

[0065] After passing through the optical fiber channel, the received signal can be expressed as:

[0066]

[0067] In the formula (h i ,w) is the channel response of the ith signal in the time domain, (H i ,W) is the channel response in the frequency domain. After receiving the signal, channel estimation and equalization are performed by comparing the sent and received training sequences. The specific equalization process can be expressed as:

[0068]

[0069] Among them, s TS and s′ TS denote the training sequences sent and received respectively, s and s′ denote the signals before and after equalization respectively, and mean(·) is the equalization function.

[0070] Specifically, refer to Figure 6 As shown, the process of the region determination method in step 6 includes:

[0071] Step 61: After channel equalization, the high-power QPSK signal is firstly subjected to regional decision, and it is determined that the superimposed signal corresponds to symbols "00", "01", "10" and "11" in the four quadrants (1, 2, 3, 4) of the constellation point, respectively;

[0072] Step 62: In the decision process of the low-power QPSK signal, the outermost constellation point is determined to be "11" and the innermost constellation point is determined to be "00";

[0073] Step 63: The remaining symbols are based on Figure 5 The corresponding coordinate area is determined in order to obtain an accurate low-power demodulated signal.

[0074] Unlike the SIC used in traditional NOMA systems, the demodulation of high-power and low-power signals is completely unrelated when executing the regional decision method. Moreover, this regional decision-based receiver does not require the complex iterative calculation process of the traditional SIC receiver, which can effectively improve the system's working efficiency and avoid the continuous propagation of errors.

[0075] Embodiment 4

[0076] Reference Figure 7 As shown, a bit-level communication probability shaping system for NOMA includes a transmitter 1, a receiver 2, a coupler 3, an optical fiber access device 4, a multi-core optical fiber 5 and an optical fiber output device 6. The transmitter 1 includes a waveform generator 11, an electrical amplifier 12, a modulator 13, an optical fiber amplifier 14 and a laser 15. The waveform generator 11 is connected to the modulator 13 through the electrical amplifier 12, the output end of the laser 15 is connected to the input end of the modulator 13, and the output end of the modulator 13 is connected to the input end of the optical fiber amplifier 14; the multi-core optical fiber 5 in this embodiment uses a seven-core optical fiber, the waveform generator 11 uses an arbitrary waveform generator in the prior art, and the modulator 13 uses a Mach-Zehnder modulator (MZM).

[0077] The receiver 2 includes an adjustable optical attenuator 21, a photodiode 22 and a mixed signal oscilloscope 23, wherein the output end of the adjustable optical attenuator 21 is connected to the input end of the photodiode 22, and the output end of the photodiode 22 is connected to the input end of the mixed signal oscilloscope 23;

[0078] The output end of the optical fiber amplifier 14 is connected to the optical fiber access device 4 through the coupler 3 , and the optical fiber access device 4 is connected to the optical fiber output device 6 through the multi-core optical fiber 5 , and the output end of the optical fiber output device 6 is connected to the input end of the adjustable optical attenuator 21 .

[0079] A bit-level communication probability shaping method for NOMA uses the above-mentioned bit-level communication probability shaping system for NOMA to verify the bit-level communication probability shaping, and the operation method flow is as follows:

[0080] Step 71: At the transmitting end, in the offline DSP, two different users generate random bit streams for transmission;

[0081] Step 72: Both sets of transmission data are modulated into QPSK-OFDM signals, a cyclic prefix is ​​added to each symbol as a guard band, different power allocations are performed on the two QPSK-OFDM signals, and then the two signals are superimposed to obtain a NOMA-OFDM signal. At this time, the multiplexed signal contains the transmission data of two users;

[0082] Step 73: After the NOMA-OFDM signal passes through the waveform generator 11, the modulated signal is converted into a radio frequency signal. In order to ensure the reference of the comparative experiment, different sampling rates of the arbitrary waveform generator are set when different k values ​​are taken to ensure that the net rate of each group of experiments is the same;

[0083] Step 74: The radio frequency signal passing through the electrical amplifier 12 is injected into the modulator 13 for photoelectric modulation;

[0084] Step 75: The optical signal carrying information is divided into seven beams through the optical fiber amplifier 14 and the coupler 3 in sequence, and then merged into the 2-kilometer-long seven-core optical fiber through the optical fiber access device 4;

[0085] Step 76: At the receiving end, the optical signal is output by the optical fiber output device 6 and then injected into the adjustable optical attenuator 21 to adjust the received optical power;

[0086] Step 77: Implement photoelectric signal conversion through the photodiode 22, and implement analog-to-digital conversion through the mixed signal oscilloscope 23, and finally obtain a received signal;

[0087] Step 78: Input the received multiplexed signal into the offline DSP module, and restore the NOMA-OFDM signal after performing channel estimation and equalization on the received signal;

[0088] Step 79: Perform regional decision and low-power signal de-BWDM on the NOMA-OFDM signal in turn, and restore the QPSK-OFDM signals of the two users in turn.

[0089] In order to verify the performance of the improved BWDM in this embodiment, the change of the generalized mutual information (GMI) of the AWGN channel under different signal-to-noise ratios (SNR) is simulated in digital signal processing.

[0090] Reference Figure 8 As shown in the figure, when k is 2, 4, and 6 respectively, the information entropy of the superimposed 16QAM is 3.91, 3.95, and 3.98 bits / symbol respectively, and its GMI changes are shown in the figure. The specific GMI calculation formula is as follows:

[0091]

[0092] Where, χ represents the symbol set of the 16QAM constellation diagram, bk,i is the ith bit of the kth transmitted symbol. The results show that for the AWGN channel, the present invention can generate significant shaping gain in the low signal-to-noise ratio region.

[0093] Furthermore, the present invention also verifies through experiments the performance improvement brought by the proposed system through probabilistic shaping.

[0094] Reference Fig. 9 As shown in the figure, the bit error rate of the multiplexed signal of the present invention and uniform 16QAM and non-probability shaping varies with the received optical power at different received optical power (ROP) levels. Here, the bit error rate is taken as the average bit error rate of the high-power and low-power signals, and the value of k is 2 by default. The experimental results show that compared with the multiplexed signal of uniform 16QAM and non-probability shaping, the multiplexed signal of the scheme proposed in the present invention has higher receiving sensitivity. Specifically, at the same net rate, at 3.8×10 -3 Under the hard decision FEC bit error rate threshold of , the present invention can obtain receiver sensitivity gains of 0.4dB and 0.33dB respectively.

[0095] In addition, further experiments verified the performance differences of the BWDM scheme under different values ​​of k. The experiments were carried out when k=2, 4, and 6 respectively.

[0096] Reference Fig.10 As shown in the figure, when ROP = -10dBm, the constellation diagram and the probability distribution of constellation points in three cases are shown in the figure on the left side (a) of the figure above. In order to reduce the bit error rate of the system and maximize the performance of the receiver, the value of k can be reduced as much as possible in the improved BWDM solution, and the original bits can be grouped more finely. Fig.10 As shown in the graph on the upper right (b), when k = 2, the number of original bit groups increases and the probability shaping degree is the highest. However, as the value of k decreases, the number of MSBs embedded in the data stream as label bits increases significantly, which also introduces a certain degree of data redundancy into the system. By calculation, when k = 2, 4, and 6, the proportion of redundant bits is 16.7%, 10.3%, and 7.1%, respectively. Therefore, according to different system application requirements, the value of k can be changed to adapt to more efficient or more accurate communication data transmission scenarios, which further proves that the present invention has high flexibility and applicability.

[0097] In addition, the bit-level probability shaping scheme adopted by the present invention only requires a small number of configurable logic modules in the hardware implementation process, and does not require any block random access memory. Through calculation, it can be concluded that compared with the distribution matching, prefix-free distribution matching and hierarchical distribution matching schemes, the present invention can save 99.30%, 78.03% and 43.39% of the configurable logic module utilization respectively. These results show that this method supports more flexible operations and is seamlessly compatible with subsequent bit-level processes to meet the specific requirements of the system in different situations, which further proves the advantage of the framework in resource utilization.

[0098] In addition, it should be understood that although the present specification is described according to implementation modes, not every implementation mode contains only one independent technical solution. This description of the specification is only for the sake of clarity. Those skilled in the art should regard the specification as a whole. The technical solutions in each embodiment may also be appropriately combined to form other implementation modes that can be understood by those skilled in the art.

Claims

1. A bit-level communication probability shaping method for NOMA, characterized by: The following steps are included: Step 1: The data transmitted by two different users are modulated into QPSK-OFDM signals; Step 2: For one of the QPSK-OFDM signals, implement symbol probability transformation through improved BWDM; Step 3: Power allocation and superposition of two QPSK-OFDM signals to form a NOMA-OFDM signal, where the signal in step 2 is allocated with low power; Step 4: Transmit the NOMA-OFDM signal through optical fiber; Step 5: After receiving the signal, perform channel estimation and equalization on the signal; Step 6: Perform regional decision on the equalized signal, and restore the high-power QPSK-OFDM signal and the low-power QPSK-OFDM signal after symbol probability transformation respectively through the regional decision method of high-power and low-power signals; Step 7: Get the recovered low-power QPSK-OFDM signal by de-BWDM; The way to modulate the transmission data of two different users in step 1 into QPSK-OFDM signals is that the two OFDM signals with different powers are both QPSK modulated, and the superimposed signal is further mapped to 16QAM. On the basis of retaining the mapping rule of the second quadrant, the low-power signals of the first, fourth and third quadrants are successively superimposed by rotating 90°, 180° and 270° clockwise; the signal after power division multiplexing can be mathematically expressed as: In the formula, And s i and p i represents the signal and signal power of the i-th user; The process of implementing symbol probability transformation of the QPSK-OFDM signal in step 2 through improved BWDM includes: Step 21: Divide the original bits of the M-bit serial low-power QPSK-OFDM signal into two parallel bit sequences (G1 and G2) of length M / 2 through cross mapping; Step 22: Divide G2 into k-bit bit combinations and weight them according to their original elements; add an extra bit as an extra significant weighted bit (MSB) in each bit group as a decision weight; in the bit weighted decision process, if the number of "0" in the k-bit is less than k / 2, the corresponding MSB is defined as "1"; Step 23: Flip the remaining bits in the bit group bit by bit, otherwise, set the MSB to "0" and keep the original bit unchanged; the bit-by-bit inversion operation is a key step in BWDM and is used to increase the probability of the "0" bit in the sequence G2 before QPSK mapping; Step 24: After bit flipping, the MSB is integrated and placed as a label in the sequence G2 data stream. After G1 and G2 perform parallel to serial (P / S) conversion, the data stream is modulated into QPSK-OFDM format.

2. The bit-level communication probability shaping method for NOMA according to claim 1 is characterized in that: The NOMA-OFDM signal production process in step 3 includes, Step 31: For two signals with different powers, the initial binary bit data is mapped to multiple subcarriers of the QPSK signal through serial-to-parallel conversion (S / P) to obtain an OFDM signal in the frequency domain; Step 32: directly perform an inverse fast Fourier transform (IFFT) on one of the low-power OFDM signals to convert it from the frequency domain to the time domain; first insert a training sequence (TS) into the high-power OFDM signal for channel estimation and equalization at the receiving end, and then perform an inverse fast Fourier transform (IFFT) to convert it from the frequency domain to the time domain; Step 33: Add a cyclic prefix (CP) in the time domain of the two OFDM signals to avoid inter-symbol interference; Step 34: Perform parallel-to-serial conversion on the two signals respectively, perform power allocation, and obtain the NOMA-OFDM signal through superposition coding.

3. The bit-level communication probability shaping method for NOMA according to claim 2 is characterized in that: The signal channel estimation and equalization process in step 5 includes, After passing through the optical fiber channel, the received signal can be expressed as: In the formula (h i ,w) is the channel response of the ith signal in the time domain, (H i ,W) is the channel response in the frequency domain. After receiving the signal, channel estimation and equalization are performed by comparing the sent and received training sequences. The specific equalization process can be expressed as: Among them, S TS and S′ TS denote the training sequences sent and received respectively, s and s′ denote the signals before and after equalization respectively, and mean(·) is the equalization function.

4. The bit-level communication probability shaping method for NOMA according to claim 3 is characterized in that: The process of the region determination method in step 6 includes: Step 61: After channel equalization, perform regional decision on the high-power QPSK signal to determine that the superimposed signal corresponds to symbols "00", "01", "10" and "11" in the four quadrants (1, 2, 3, 4) of the constellation point, respectively; Step 62: In the decision process of the low-power QPSK signal, the outermost constellation point is determined to be "11" and the innermost constellation point is determined to be "00"; Step 63: The remaining symbols are determined according to the corresponding coordinate areas in FIG. 5 , so as to obtain an accurate low-power demodulated signal.

5. A bit-level communication probability shaping system for NOMA, used to verify any one of the bit-level communication probability shaping methods for NOMA in claims 1-4, characterized in that: It includes a transmitter 1, a receiver 2, a coupler 3, an optical fiber access device 4, a multi-core optical fiber 5 and an optical fiber output device 6. The transmitter 1 includes a waveform generator 11, an electrical amplifier 12, a modulator 13, an optical fiber amplifier 14 and a laser 15. The waveform generator 11 is connected to the modulator 13 through the electrical amplifier 12, the output end of the laser 15 is connected to the input end of the modulator 13, and the output end of the modulator 13 is connected to the input end of the optical fiber amplifier 14. The receiver 2 comprises an adjustable optical attenuator 21, a photodiode 22 and a mixed signal oscilloscope 23, wherein the output end of the adjustable optical attenuator 21 is connected to the input end of the photodiode 22, and the output end of the photodiode 22 is connected to the input end of the mixed signal oscilloscope 23; The output end of the optical fiber amplifier 14 is connected to the optical fiber access device 4 through the coupler 3 , and the optical fiber access device 4 is connected to the optical fiber output device 6 through the multi-core optical fiber 5 , and the output end of the optical fiber output device 6 is connected to the input end of the adjustable optical attenuator 21 .

6. A NOMA-oriented bit-level communication probability shaping method, using the NOMA-oriented bit-level communication probability shaping system in claim 5, characterized in that: It is used to verify the bit-level communication probability shaping, and its operation method flow is as follows: Step 71: At the transmitting end, in the offline DSP, two different users generate random bit streams for transmission; Step 72: Both sets of transmission data are modulated into QPSK-OFDM signals, a cyclic prefix is ​​added to each symbol as a guard band, different power allocations are performed on the two QPSK-OFDM signals, and then the two signals are superimposed to obtain a NOMA-OFDM signal. At this time, the multiplexed signal contains the transmission data of two users; Step 73: After the NOMA-OFDM signal passes through the waveform generator 11, the modulated signal is converted into a radio frequency signal. In order to ensure the reference of the comparative experiment, different sampling rates of the arbitrary waveform generator are set when different k values ​​are taken to ensure that the net rate of each group of experiments is the same; Step 74: The radio frequency signal passing through the electrical amplifier 12 is injected into the modulator 13 for photoelectric modulation; Step 75: The optical signal carrying information is divided into multiple beams through the optical fiber amplifier 14 and the coupler 3 in sequence, and then merged into the 1 to 3 km long multi-core optical fiber 5 through the optical fiber access device 4; Step 76: At the receiving end, the optical signal is output by the optical fiber output device 6 and then injected into the adjustable optical attenuator 21 to adjust the received optical power; Step 77: Implement photoelectric signal conversion through the photodiode 22, and implement analog-to-digital conversion through the mixed signal oscilloscope 23, and finally obtain a received signal; Step 78: Input the received multiplexed signal into the offline DSP module, and restore the NOMA-OFDM signal after performing channel estimation and equalization on the received signal; Step 79: Perform regional decision and low-power signal de-BWDM on the NOMA-OFDM signal in turn, and restore the QPSK-OFDM signals of the two users in turn.

Images