Optical fiber-visible light converged communication system for transmitting end signal reverse probability shaping

Abstract

The invention belongs to the technical field of optical communication, and particularly discloses an optical fiber-visible light converged communication system for transmitting end signal reverse probabilistic shaping. According to the method, carrier waves are injected into a base band, a single side band signal is formed, a KK receiving scheme meeting the minimum phase condition is combined with an RPS strategy of distributing constellation probability to an outer ring, efficient and steady recovery of a mixed link receiving signal under the conditions of low signal-to-noise ratio and limited peak power is achieved, and the method has the advantages of being simple in structure and convenient to operate. And the phase reconstruction reliability is improved by setting a proper carrier signal power ratio. The method overcomes the defects of insufficient stability of peak constraint and phase recovery when strong carrier waves and low signal-to-noise ratio coexist in a traditional direct detection link and high peak-to-average power ratio and aggravated mutual information degradation of conventional probability shaping under the low signal-to-noise ratio, has higher noise immunity and spectral efficiency, and can be applied to the field of wireless communication. The method is a feasible scheme for a 6G indoor optical wireless access and optical fiber wireless optical integrated network.

Description

Technical Field

[0001] This invention relates to the field of optical wireless and wired converged communication technology, specifically to an optical fiber-visible light converged communication system and method that combines inverse probability shaping and a KK receiver. Background Technology

[0002] With the rapid development of services such as 6G and the Internet of Things (IoT), future communication systems need to simultaneously meet the demands for high-capacity connectivity, ultra-low latency, and highly flexible access. Current fiber optic communication technology, with its ultra-high bandwidth and low-loss characteristics, remains the backbone for long-distance, high-capacity information transmission. However, at the "last mile" access level, fiber optic deployment is limited by complex construction, insufficient flexibility, and high costs.

[0003] Therefore, the fiber-visible light converged communication system has emerged as a novel architecture combining "fiber optic backbone transmission capabilities" with "wireless access flexibility." This system utilizes fiber optics for long-distance, high-speed transmission, and then uses visible light communication (VLC) for short-distance free-space transmission, thereby constructing a low-cost, high-capacity integrated optical-wireless access network. Compared to traditional radio frequency communication, VLC offers significant advantages such as unlicensed spectrum, resistance to electromagnetic interference, low power consumption, and high spatial multiplexing, making it particularly suitable for high-density indoor scenarios and multi-terminal access requirements.

[0004] However, current optical wireless communication systems still predominantly employ an intensity modulation / direct detection (IM / DD) architecture at the physical layer. While this architecture simplifies system costs, it lacks I / Q modulators that directly support complex modulation in the visible light band, hindering the efficient transmission of high-order complex signals. Furthermore, IM / DD systems suffer from significant signal distortion and phase recovery difficulties under low received optical power (ROP) and high-noise environments, severely limiting system performance. In response, the KK receiver, proposed in recent years, offers a low-complexity alternative to coherent detection for direct detection systems. Based on the Hilbert transform principle, the KK receiver can reconstruct the complex optical field from the intensity signal, thereby recovering phase information without the need for a local oscillator and optical mixer. Its core principle is satisfying the minimum phase condition, i.e., the signal must be in single-sideband form. The KK receiver ensures this condition by inserting a strong carrier wave into the signal, but this simultaneously increases the system's peak power, potentially introducing nonlinearity.

[0005] On the other hand, probabilistic shaping, by adjusting the symbol probability distribution at constellation points to approximate channel capacity, is widely used in optical fiber communication to improve spectral efficiency and information rate. Traditional probabilistic shaping is based on positive shaping parameters, which increase the probability of inner-loop symbols and decrease the probability of outer-loop symbols, thereby reducing average power and improving noise immunity. However, under low signal-to-noise ratio (SNR) or peak-limited conditions, the performance of classical probabilistic shaping degrades because the inter-symbol distance decreases, leading to reduced noise immunity. Summary of the Invention

[0006] The technical problem to be solved by this invention is to propose an optical fiber-visible light fusion communication system based on inverse probability shaping and KK receiver, which can achieve highly reliable transmission of complex modulation signals in scenarios with low ROP, low SNR and limited peak power.

[0007] To address the aforementioned technical problems, this invention provides a fiber-visible light fusion communication system with transmitter signal inverse probability shaping, comprising a transmitter and a fiber-visible light fusion communication link connecting the transmitter and the receiver. The transmitter is used to inject a carrier wave into the baseband to form a single-sideband signal that satisfies the KK algorithm for reception; before carrier injection, reverse probability shaping is used to adjust the symbol probability distribution of the baseband bit sequence mapping constellation points so that the probability is distributed outwards. The receiving end includes a receiver based on the KK algorithm and a receiving end digital signal processor; The fiber-visible light fusion communication link connects the transmitter and the receiver.

[0008] The reverse probability shaping adjustment of the baseband bit sequence mapping constellation points uses the following formula: in, Dots for constellations. This is the inverse shaping coefficient, and its value range is... .

[0009] The fiber-visible light fusion communication link includes a fiber optic transmission link, a visible light free-space transmission link, and an optical-electrical-optical signal conversion device. The fiber optic transmission link is used to receive the single-sideband signal from the transmitting end and transmit the single-sideband signal over a long distance. The visible light free-space transmission link receives the signal from the fiber optic transmission link through the optical-electrical-optical signal conversion device and transmits the signal over a short distance in indoor visible light free space. The optical-electrical-optical signal conversion device is used for optical / electrical signal conversion during the fiber-visible light fusion communication process. The transmitting end includes a digital signal processor, which performs the following sequentially: generating a baseband signal bit sequence; mapping the baseband signal bit sequence onto constellation points of M-order orthogonal amplitude modulation; adjusting the transmission probability of each constellation point using inverse probability shaping; upsampling the baseband signal after inverse probability shaping; applying root-raised cosine shaping filter to form a complex baseband signal; introducing a carrier into the complex baseband signal to generate a single-sideband carrier signal that meets the requirements of the KK algorithm for reception; upconverting and modulating the carrier signal; and separating in-phase orthogonal signals. The carrier signal power ratio is between 8dB and 14dB to ensure that the minimum phase condition is met.

[0010] The main optical path of the optical fiber transmission link is a single-mode optical fiber with a maximum length of 20 kilometers.

[0011] The visible light free space transmission link adopts free space transmission with a transmission distance of 0.5 meters to 4 meters. The beginning of the visible light free space transmission link emits light through a collimating lens, and the end receives light through a focusing lens, and direct detection is performed using a PIN photodiode.

[0012] The optical-to-electrical-to-optical signal conversion device includes a first conversion device and a second conversion device. The first conversion device includes an avalanche photodiode, a first electrical amplifier, a DC biaser, and a visible light band laser diode. The avalanche photodiode is connected to the end of the optical fiber transmission link to realize optical-to-electrical conversion. The output of the optical fiber passes through the avalanche photodiode and then through the first electrical amplifier to drive the visible light band laser diode to realize electrical-to-optical conversion. The converted optical signal is input to the beginning of the visible light free space transmission link.

[0013] The second conversion device includes a photodiode and a second electrical amplifier. The photodiode is connected to the end of the visible light free space transmission link to achieve photo-to-electric conversion. The electrical signal is then amplified by the first electrical amplifier and output to the system receiver. The receiver based on the KK algorithm is used to perform complex field reconstruction on the received signal by square root extraction, logarithmic operation, Hilbert transform, and exponential recovery.

[0014] The receiver digital signal processor implements: waveform resampling and downsampling, KK field reconstruction, frequency downconversion, recursive least squares equalization, Gram-Schmidt orthogonal projection I / Q imbalance correction, constant modulus algorithm equalization, carrier phase estimation, and decision-oriented least mean square equalization and inverse probabilistic shaping constellation demapping.

[0015] This invention relates to a heterogeneous communication network system integrating fiber optic and visible light communication based on reverse probabilistic shaping (RPS) and a KK receiver. The system employs an RPS strategy, adjusting the probability distribution of constellation points by applying a negative shaping coefficient. This increases the probability of high-energy outer ring symbols, and the negative shaping factor improves average power and reduces PAPR, thereby reducing peak power issues caused by KK carrier insertion, improving nonlinear distortion, and enhancing signal discrimination and phase recovery accuracy under low SNR conditions. Experimental results show that, in a 20km fiber optic and 0.5–4m visible light link environment, the synergistic effect of RPS and KK detection significantly reduces PAPR by 0.45dB compared to unshaped 16QAM, effectively lowering the bit error rate. Specifically, it achieves a net data rate of 10.36Gbps at a 0.5m free-space transmission distance and maintains 8.7Gbps at a 4m free-space transmission distance, outperforming traditional QAM modulation. Compared to CPS, RPS exhibits slower GMI and NGMI degradation and maintains a higher information rate in low ROP scenarios. This system effectively solves the problems of peak power limitation and phase recovery difficulty under low SNR in visible light links, and realizes a highly robust and highly spectral efficient integrated optical wired-wireless communication architecture, which is suitable for next-generation 6G broadband access and optical-wireless converged networks. Attached Figure Description

[0016] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0017] Figure 1 This is a diagram of the main experimental setup for the fiber-visible light fusion communication system based on inverse probability shaping and a KK receiver, as described in this invention.

[0018] Figure 2 This describes the processing procedure of the digital signal processor (DSP) at the transmitting end of the system of the present invention.

[0019] Figure 3 This describes the processing procedure of the digital signal processor (DSP) at the receiving end of the system of the present invention.

[0020] Figure 4 These are the calculation steps of the KK algorithm used in this invention.

[0021] Figure 5 The graph shows the performance of the present invention at low peak-to-average power ratio (PAPR) and error vector amplitude (EVM) under different carrier signal power ratio (CSPR) settings, and compares it with the classical method. Figure 5 The three curves in (a) are the PAPR versus CSPR curves for three 16th-order orthogonal amplitude modulation cases: classic 16QAM, conventional probabilistic shaping CPS-16QAM, and the inverse probabilistic shaping RPS-16QAM of the present invention. Figure 5 (b) in the middle is with Figure 5 The curve of EVM versus CSPR corresponding to (a) in the figure.

[0022] Figure 6 This is a comparison chart of the generalized mutual information (GMI) and the normalized generalized mutual information (NGMI) of the method of this invention and the classical method under the same modulation order and different ROP conditions. Figure 6 (a) in the figure represents back-to-back (B2B) communication; (b) in the figure represents 20 km of optical fiber + 1 m of VLC free space transmission.

[0023] Figure 7 This is a comparison chart of GMI / NGMI using the method of this invention and the classical method under the same source entropy constraint (H=4) but with different modulation orders. Figure 7 (a) is a comparison chart of NGMI under B2B communication; (b) is a comparison chart of NGMI under 20 km fiber + 1 m VLC free space transmission.

[0024] Figure 8 The graph shows the end-to-end system performance under varying free-space VLC transmission distances and a comparison with classical methods. Figure 8 In (a), the maximum achievable NDR and corresponding NGMI are given under the condition of distances of 0.5–4 m; Figure 8 (b) in the table shows the results and comparison of BER and EVM as a function of distance, corresponding to (a) in 8. Detailed Implementation

[0025] The invention will be further described below with reference to the system and results of specific experiments.

[0026] The construction steps of this invention are as follows: (1) Constructing a transmitter DSP module based on inverse probabilistic shaping The transmitter first randomly generates a pseudo-random bit sequence (PRBS), maps it to M-order quadrature amplitude modulation (QAM) constellation points, and adjusts the transmission probability of each constellation point according to a reverse probability shaping strategy. The symbol probability distribution is as follows: in, Dots for constellations. This is the inverse shaping coefficient, and its value range is... The optimal value is Negative values ​​indicate an increased probability of high-energy symbols appearing in the outer ring, thereby increasing the average power, reducing PAPR, and enhancing symbol distinguishability.

[0027] After upsampling and RRC shaping filtering, a complex baseband signal is formed. Then, a carrier wave is introduced into the baseband to ensure the minimum phase condition and generate an SSB signal, which is finally output to an arbitrary waveform generator (AWG) for digital-to-analog conversion.

[0028] (2) Construct a long-distance transmission module based on fiber optic links The I / Q electrical signal output by the AWG drives a 1550nm tunable laser to be optically modulated by an IQ modulator. The output optical signal is amplified by an erbium-doped fiber amplifier (EDFA) and then transmitted through a 20km SSMF.

[0029] Optical-to-electrical (O / E) conversion is achieved at the fiber optic end using an avalanche photodiode (APD), converting the optical signal into a baseband electrical signal. The linear impairments introduced in this process include attenuation and dispersion, which will be compensated for later in the receiver's DSP.

[0030] (3) Construct a visible light link with an optical-electrical-optical conversion module The electrical signal output by the APD is amplified by the EA and drives a 660nm photodiode LD to form a visible light emitter, realizing electro-optical E / O conversion. This optical-electrical-optical cascade architecture achieves seamless integration of fiber optic backbone and visible light wireless access.

[0031] The visible light link adopts a free-space transmission structure. A collimating lens is configured at the beginning of the visible light link, and a focusing lens is configured at the receiving end. The transmission distance is adjustable within the range of 0.5–4m. A PIN photodiode is used at the end for direct detection, and the amplified electrical signal is output to a high-speed oscilloscope for sampling.

[0032] (4) Construct a complex field recovery module based on the KK algorithm The receiver performs complex field reconstruction on the detected intensity signal and the photocurrent signal. With modulated signal The relationship is: in, Let be the responsivity of the photodetector, and t be the signal period time slot. If the input signal satisfies the minimum phase condition, the phase information can be reconstructed from the KK relation: in, Represents Cauchy's principal value integral. The instantaneous phase is given. This integral is calculated using the Hilbert transform implemented by the Fast Fourier Transform (FFT).

[0033] The reconstructed signal can be further reconstructed as: This process can reconstruct complex amplitude from intensity measurements without requiring a local oscillator and optical mixer, achieving phase information recovery under direct detection conditions. To meet the minimum phase condition, the system CSPR is controlled at approximately 12 dB.

[0034] (5) Construct the receiver DSP post-processing module The signal recovered by KK reconstructed field undergoes the following processes in the receiver DSP: resampling and downsampling, frequency downconversion, recursive least squares (RLS) equalization, Gram-Schmidt orthogonalization (GSOP) I / Q imbalance correction, constant modulus (CMA) equalization, carrier phase compensation (CPE) and decision-guided least mean square (DDLMS) equalization. Finally, the bit stream is recovered by RPS-QAM demapping, and the bit error rate (BER) is calculated by comparing the transmitted and received bit streams.

[0035] This invention, using a 20km fiber optic cable and a 0.5–4m visible light link, leverages the synergistic effect of RPS and KK detection to significantly reduce PAPR (Probabilistically Shaped Precision Rate), achieving a 0.45dB reduction compared to signals without probabilistic shaping, thus effectively lowering the bit error rate. Specifically, it achieves a net data rate of 10.36Gbps over a 0.5m free-space transmission distance, a 0.6Gbps increase compared to traditional probabilistic shaping schemes.

[0036] To verify the proposed scheme, this invention experimentally demonstrated a fiber-visible light fusion communication system employing probabilistic shaping and KK detection. (1) Experimental system setup The main structure of the system is as follows Figure 1 As shown.

[0037] In the DSP section of the transmitter, such as Figure 2 As shown, a PRBS signal of 16,384 symbols is first generated and modulated onto a PS-QAM constellation diagram. The symbol rate of the signal is 2 Gbaud. The signal is then upsampled by 8 times and filtered using an RRC pulse shaping filter with a roll-off factor of 0.03. A carrier required by the KK receiver is added, with the carrier-to-signal power ratio ranging from 8 dB to 14 dB to ensure the minimum phase condition is met. The pulse-shaped I / Q components are combined with the carrier signal, with the carrier frequency offset by 0.01 times relative to the signal bandwidth, thus forming an SSB signal. The generated signal is then up-converted and separated into in-phase (I) and quadrature (Q) components. The separated digital signal is then processed by... Figure 1 The AWG (Keysight M8195A, sampling rate 65GSa / s) output has an effective sampling rate of 16 GSa / s and is modulated onto the optical carrier emitted by a 1550nm tunable laser source through an IQ modulator.

[0038] The modulated optical signal is amplified by an EDFA (maximum transmit power 10dBm) and transmitted over a 20km SSMF link. The ROP (Resonance Optical Output) is adjusted using a variable optical attenuator (VOA, attenuation range 0–40dB). At the fiber output, an APD converts the optical signal into an electrical signal. This converted electrical signal is amplified by an EA (Electronic Amplifier) ​​with a bandwidth of 4GHz and a gain of 30dB, driving a 660nm wavelength LD (Light Detector), which serves as the transmitter in the VLC (Visible Light Array) front-end. The VLC link is a free-space transmission link with an adjustable transmission distance ranging from 0.5m to 4m. The experimental system supports a direct transmission distance of 2m, and optionally, a reflector can be inserted to reflect the beam, achieving a maximum visible light transmission distance of 4m. A collimating lens (CL) is used at the transmitter to emit the beam, and a focusing lens (FL) is used at the receiver to collect the optical signal. The received VLC optical signal is converted into an electrical signal by a PIN photodiode, then amplified by an EA and sampled in real time by a high-speed oscilloscope (OSC, Agilent DSA-X 92004A, sampling rate 80 GSa / s).

[0039] exist Figure 3 In the DSP of the receiver shown, the acquired waveform is first resampled to match the effective sampling rate of the AWG and oscilloscope. Then, the signal is downsampled by a factor of 8 to match the processing bandwidth. Next, applications such as... Figure 4 The KK algorithm shown reconstructs the complex optical field from the detected intensity signal. This process includes sequentially performing square root, logarithmic, Hilbert transform, and exponential recovery steps. After the complex optical field is recovered, the signal is down-converted from the intermediate frequency (IF) to the baseband. Then, the RLS recursive least squares equalization algorithm is used to compensate for channel impairments. Next, the GSOP orthogonalization algorithm is used to correct I / Q imbalance. Then, the CMA constant modulus algorithm is used for further equalization to eliminate residual distortion. After CMA, carrier phase compensation (CPE) is performed to correct the phase rotation introduced by blind equalization and improve the system's tolerance to phase noise. Subsequently, the DDLMS decision-guided least mean square algorithm is used to further optimize symbol decision, improving overall system performance. Finally, PS-QAM demapping is performed to recover the transmitted data.

[0040] (2) Experimental results and analysis To investigate the impact of different probabilistic shaping strategies on the performance of a KK receiver-based fiber-visible light fusion system, the effect of CSPR on key transmission parameters was first analyzed. This experiment aims to explore the influence of different shaping strategies on PAPR and EVM under different CSPR conditions, as these factors directly determine the phase recovery performance of the KK receiver. Figure 5Figures (a) and (b) show the PAPR and EVM performance of unshaped 16QAM, CPS-16QAM, and RPS-16QAM under different CSPR values, respectively. The experiments were conducted after 20km of fiber transmission and a 1m free-space VLC link to simulate a real-world hybrid transmission scenario. Figure 5 As shown in result (a), RPS achieves the lowest PAPR across the entire CSPR range, while CPS actually achieves a higher PAPR than unshaped 16QAM. This is because CPS concentrates the symbol probability at the center of the constellation diagram, leading to a reduction in average power. Regarding EVM performance, such as... Figure 5 As shown in (b), RPS-16QAM exhibits the lowest EVM and the most stable performance under different CSPRs, while CPS performs poorly under all conditions. The carrier added to meet the minimum phase requirement of the KK receiver keeps the peak power almost unchanged, but further increases the PAPR under CPS conditions. The high EVM of CPS is mainly due to its high PAPR, making the signal more susceptible to clipping and amplifier nonlinear distortion, thus deteriorating the overall signal quality. Therefore, RPS is more stable in maintaining signal quality. According to the EVM measurement results, the system achieves its best performance at CSPR=12dB, at which point all shaping methods are in their optimal state, with RPS exhibiting the lowest EVM, indicating that the fiber-visible light fusion system using RPS and KK detection has the best overall performance. The constellation diagram at this CSPR also shows that RPS has the clearest symbol clustering and the least distortion, meeting the characteristics of the KK receiver. Quantitatively, at CSPR=12dB, RPS is 0.45dB lower than PAPR of unshaped 16QAM, while CPS is 0.21dB higher, further demonstrating that RPS has better nonlinear tolerance under high peak power conditions.

[0041] Figure 6This study demonstrates the performance of NGMI and GMI under different probabilistic shaping strategies at varying received optical power (ROP). Experiments were conducted in (a) B2B and (b) hybrid links of "20km fiber + 1m free space" to compare the performance of CPS and RPS at the same QAM order. The results show a high degree of consistency between the trends under the two transmission conditions, validating the reliability of the experimental conclusions. As ROP increases, the system SNR improves, and NGMI gradually approaches 1. When ROP is -20dBm, all modulation formats approach ideal performance; however, when ROP decreases, SNR deteriorates, leading to a decrease in NGMI. Under the same source entropy conditions, the curve of RPS-16QAM is consistently higher than that of CPS-16QAM, indicating that RPS exhibits slower NGMI degradation and superior performance under low-power conditions. When ROP is below -26dBm, even with higher source entropy, RPS still outperforms CPS with lower source entropy, demonstrating that RPS has stronger signal discrimination after KK detection, especially in low-power and distortion-prone environments. The GMI results also show a similar trend: as ROP increases, GMI approaches the modulation entropy, resulting in minimal information loss; while when ROP decreases, CPS's GMI decreases faster, while RPS's decreases more slowly, demonstrating its better noise and distortion resistance. This performance improvement can be attributed to the complementary characteristics between the RPS shaping method and the KK receiver's phase recovery process: in low ROP scenarios, the signal amplitude is enhanced relative to carrier fluctuations, making phase recovery more sensitive to noise; RPS increases the probability distribution of high-energy symbols in the outer loop, improving the average signal power without changing the peak power, thereby enhancing the phase extraction SNR and symbol discriminability of the KK algorithm. In contrast, CPS compresses the constellation distribution, resulting in a weaker signal envelope, making it more vulnerable to noise and nonlinear distortion. Therefore, under the same source entropy conditions, RPS can maintain a clearer signal structure and achieve more reliable decoding performance.

[0042] Figure 7Further analysis was conducted on the GMI and NGMI performance of different QAM orders and probabilistic shaping strategies under various ROP conditions, under the same source entropy constraint (H=4). The 64QAM signal was shaped to maintain a source entropy of 4, consistent with 16QAM, to ensure fair comparison. Results show that as ROP decreases, the system SNR decreases, and the performance gap between different schemes gradually widens. In the high ROP region, due to the larger constellation point spacing and simpler mapping, the unshaped 16QAM exhibits the highest NGMI and GMI; while CPS-64QAM is extremely sensitive to ROP changes, with both indicators decreasing significantly at low ROP. When ROP drops below -22dBm, the NGMI of RPS-64QAM exceeds that of the unshaped 16QAM, and its GMI is comparable to 16QAM and significantly better than CPS-64QAM. As ROP continues to decrease, the performance gap between RPS and other schemes further widens, highlighting its significant advantages. This demonstrates that RPS maintains good symbol distinguishability and reduces the impact of noise and nonlinear distortion, even under high-order modulation and complex constellation conditions, thereby improving system robustness. GMI results further prove that RPS-64QAM maintains higher mutual information and less information loss at low ROP. Due to the dense constellation points and small symbol spacing in 64QAM, noise is particularly significant. RPS, by increasing the probability of outer-loop symbols, improves the average symbol energy and effective distance between symbols, thus achieving stronger noise immunity. Conversely, CPS concentrates inner-loop symbols, making the system more susceptible to noise and nonlinear interference. Therefore, RPS exhibits higher stability and better performance under different modulation orders and ROP conditions, verifying its applicability to high-order modulation in power-constrained VLC systems.

[0043] Figure 8 The performance variations of RPS, CPS, and unshaped QAM at different free-space VLC transmission distances are demonstrated. Figure 8Figure (a) shows the maximum achievable NDR and corresponding NGMI of the system under the condition that the BER is below the 7% HD-FEC threshold. The red curve represents the baseline performance of adaptive ordinary QAM; based on this, higher-order QAM combined with probabilistic shaping is used to achieve higher source entropy within the BER threshold. The shaping parameters of RPS-QAM and CPS-QAM are adaptively adjusted according to the transmission distance to achieve their respective optimal performance. As the transmission distance increases from 0.5m to 4m, the NDR gradually decreases due to path loss and noise. Specifically, the source entropy of RPS-QAM decreases from 5.6 to 4.7, while that of CPS-QAM decreases from 5.3 to 4.5 to maintain the highest usable rate under the bit error rate constraint. When the distance is 4m, the performance drops sharply, mainly because the beam divergence effect at long distances increases the beam spot size. Even with collimating and focusing lenses, some optical power cannot be collected by the receiving lens, resulting in significant received power loss. At the same time, the influence of environmental noise increases, thus limiting the effective ranging range of the system. Nevertheless, RPS-QAM outperforms CPS-QAM in both NDR and NGMI across all distances. At a distance of 1m, RPS reaches 10.36Gbps, approximately 0.6Gbps higher than CPS; at 4m, RPS remains above 8.7Gbps, while CPS drops below 8.3Gbps. The constellation distribution at the transmitter and receiver further illustrates that RPS symbols are primarily distributed in the outer ring, resulting in greater Euclidean distances between symbols and a lower probability of misclassification. Figure 8 The BER results shown in (b) indicate that all measurement points are below the 7% HD-FEC threshold, and the corresponding EVM results also verify this trend—RPS maintains the lowest EVM at all distances, indicating that it can better maintain signal integrity in free-space transmission. These results fully demonstrate that RPS has stable and quantifiable performance advantages over traditional probabilistic shaping in fiber-visible light fusion systems based on KK receivers, including lower PAPR, better phase recovery conditions, slower mutual information degradation rate, and higher end-to-end transmission fidelity.

[0044] Finally, it should be noted that the above specific embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A transmitting end signal reverse probability shaping optical fiber-visible light fusion communication system, comprising a transmitting end, an optical fiber-visible light fusion communication link connecting the transmitting end and a receiving end, characterized in that, the transmitting end is used for injecting a carrier wave on a baseband to form a single sideband signal satisfying KK algorithm receiving; before the carrier wave injection, a reverse probability shaping is used to adjust a symbol probability distribution of a baseband bit sequence mapping constellation point, so that the probability is distributed outwardly; the receiving end comprises a KK algorithm based receiver and a receiving end digital signal processor; the optical fiber-visible light fusion communication link connects the transmitting end and the receiving end.

2. The transmitter signal inverse probability shaping for fiber- visible light hybrid communication system according to claim 1, wherein, The reverse probability shaping is used to adjust the symbol probability distribution of the baseband bit sequence mapping constellation point, and the following formula is used: wherein, is a constellation point symbol, is a reverse shaping coefficient, taking a value in the range .

3. The transmitter signal inverse probability shaping for a fiber- visible light converged communication system of claim 2, wherein, the optical fiber-visible light fusion communication link comprises an optical fiber transmission link, a visible light free space transmission link and an optical-electric-optical signal conversion device, the optical fiber transmission link is used for receiving the single sideband signal of the transmitting end and transmitting the single sideband signal over a long distance; the visible light free space transmission link receives the signal of the optical fiber transmission link through the optical-electric-optical signal conversion device and transmits the signal over a short distance in an indoor visible light free space; the optical-electric-optical signal conversion device is used for optical / electric signal conversion in the optical fiber-visible light fusion communication process.

4. The transmit end signal inverse probability shaping for fiber- visible light converged communication system according to claim 2, wherein, The transmitting end comprises a digital signal processor, which successively performs: generating a baseband signal bit sequence, mapping the baseband signal bit sequence to an M-order quadrature amplitude modulation constellation point, adjusting a transmission probability of each constellation point by reverse probability shaping, up-sampling the baseband signal after the reverse probability shaping, using a root-raised cosine shaping filter to form a complex baseband signal, introducing a carrier wave in the complex baseband to generate a single sideband carrier wave signal satisfying KK algorithm receiving, up-conversion modulation of the carrier wave signal, and in-phase and quadrature signal separation; the carrier wave signal power ratio is valued between 8 dB and 14 dB.

5. The transmit end signal inverse probability shaping for fiber- visible light converged communication system according to claim 3, wherein, The main light path of the optical fiber transmission link is a single mode optical fiber, and the length is maximally 20 kilometers.

6. The transmit end signal inverse probability shaping for fiber- visible light converged communication system according to claim 3, wherein, The visible light free space transmission link uses free space transmission, the transmission distance is 0.5 meters to 4 meters, the beginning end of the visible light free space transmission link emits light through a collimating lens, the end emits light through a focusing lens, and direct detection is performed using a PIN photodiode.

7. The transmit end signal inverse probability shaping for fiber- visible light converged communication system according to claim 3, wherein, The optical-electric-optical signal conversion device comprises a first conversion device and a second conversion device, the first conversion device comprises an avalanche photodiode, a first electric amplifier, a direct current bias device and a visible light wave band laser diode, the avalanche photodiode is connected to the end of the optical fiber transmission link to realize optical-electric conversion, the output end of the optical fiber transmits light through the avalanche photodiode, drives the visible light wave band laser diode through the first electric amplifier to realize electric-optical conversion, and the converted optical signal is input to the beginning end of the visible light free space transmission link; the second conversion device comprises a photodiode and a second electric amplifier, the photodiode is connected to the end of the visible light free space transmission link to realize optical-electric conversion, and the amplified electric signal is output to the system receiving end through the first electric amplifier. 8.The transmitter signal inverse probability shaping for fiber-visual light fusion communication system according to claim 1, wherein, The KK algorithm-based receiver is used for square root extraction, logarithm operation, Hilbert transform and complex field reconstruction of exponential recovery on the received signal.

9. The transmit end signal inverse probability shaping for fiber- visible light converged communication system according to claim 1, wherein, The receiving end digital signal processor realizes: Waveform resampling and downsampling, KK field reconstruction, frequency down-conversion, recursive least square equalization, Gram-Schmidt orthogonal projection I / Q imbalance correction, constant modulus algorithm equalization, carrier phase estimation, decision-directed least mean square equalization and anti-probability shaping constellation demapping.

Images