Recently, neural networks have gained prominence as potent tools in the realm of channel modeling in optical wireless communication (OWC) systems. However, existing channel networks tend to ignore random intensity noise and phase noise induced by optical sources under spontaneous and excited radiation, and likewise the shot noise of the devices. They could only characterize deterministic noise such as ໿inter-symbol interference (ISI), ໿signal-to-signal beat interference (SSBI), and nonlinearities. In this paper, a novel channel simulator framework to optical wireless communication systems is proposed, which can jointly learn the deterministic noise and random noise of the real channel simultaneously. The proposed approach is a comprehensive noise joint channel estimator (CNJCE) architecture. A comprehensive comparison is carried out in both simulations, free space optical (FSO) experiments and underwater visible light communication (UVLC) experiments, including three classical deep learning algorithms, two tributaries heterogeneous neural network (TTHnet) with posteriori additive white Gaussian noise (AWGN), white Gaussian noise layer-based channel estimator (WGNCE) and conditional generative adversarial network (CGAN) based channel model. In order to represent both deterministic noise and random noise modeling capabilities, an overall estimation score is also proposed, based on frequency mismatch and random noise distribution error. The final experimental results show that CNJCE achieved the highest average estimation scores of 79.86 and 97.52 in real FSO and UVLC channels respectively. In contrast, TTHnet without AWGN or WGNCE only achieved 39.83 and 43.82. And CGAN based channel model completely lost signal characteristics, far worse than CNJCE. Through comparison of various performance, CNJCE demonstrates commendable capabilities in generalization and robustness, concurrently achieving optimal performance in channel modeling. To the best of our knowledge, this is the first time that deterministic and random noise have been fully modeled in optical wireless communication modeling. Its applicability is showcased through successful deployment in real FSO and UVLC systems, thus establishing a foundational framework for comprehensive end-to-end learning paradigms.

The low-earth-orbit (LEO) satellite communication has been proposed to support the next generation 6G space-airground integrated network owing to its large coverage area and seamless connectivity. To overcome the challenge of long-distance transmission for the free space optical (FSO) links, satellite communication systems often rely on high power transmissions to achieve high signal-to-noise ratios (SNRs) and high throughput, albeit at the expense of increased nonlinearity. In this paper, we propose a Q-learning inspired geometric shaping (QLI-GS) optimization scheme for amplitude phase shift keying (APSK) constellation to mitigate nonlinear effects and maximize the transmission distance. A pairwise optimization (PO) algorithm inspired by Q-learning is employed to achieve QLI-GS optimization for the APSK constellation according to an enhanced nonlinear interference noise model (ENINM) with effective memory mechanism, which ensures fast convergence and suppresses nonlinear impairments. Our results show the iteration time of the QLI-GS-APSK is reduced by approximately 99.89% compared to the generalized mutual information (GMI) maximized GS algorithm, and it also offers greater nonlinear tolerance than normally multi-ring APSK and traditional POAPSK. By leveraging a fiber-optical sampled bidirectional dualterminal FSO equivalent link, the QLI-GS-32APSK can experimentally achieve a data rate of 100 Gbit/s over a 230 km distance based on a single-wavelength system. A maximum distance of 580 km at a data rate of 60 Gbit/s using QLI-GS8APSK has been successfully demonstrated under 7% hard decision-forward error correction (HD-FEC) threshold of 3.8e-3, with a corresponding link budget of 62 dB. Additionally, the robustness of QLI-GS optimization under different atmospheric turbulence conditions has been further validated. We demonstrate single carrier 100Gbit/s QLI-GS-32APSK rate transmission over a 4.6 km outdoor FSO link with an Fried's coherence length of 11 cm, achieving 0.9 dB shaping gain over 32APSK. The results highlight the promising potential of the proposed QLI-GS-APSK scheme in addressing nonlinear challenges in satellite communications.

Optical encryption provides strong physical-layer security but is limited by the slow response of spatial light modulators. We propose and experimentally demonstrate a spatiotemporal noise chaffing system inspired by the “chaffing and winnowing” principle for ultrahigh-speed temporal encryption. By exploiting the symmetric spatial properties and orthogonality of conjugated orbital angular momentum (OAM) states, high-speed temporal signals (“wheat”) and spatial noise (“chaff”) are simultaneously encoded. This mechanism suppresses information leakage by degrading the temporal signal-to-noise ratio while enabling authorized recovery. Furthermore, a variable-weight multimodal OAM (VW-multimodal OAM) scheme combined with a multimodal generation neural network (MGNN) exponentially expands the key space beyond 10¹⁰. Experimentally, a record secure transmission rate of 1.25 Tbps per mode is achieved in an eight-channel wavelength-division-multiplexed coherent link. The product of rate and key space surpasses existing methods by five orders of magnitude, establishing a new photonic-security paradigm for future ultrafast and secure communication networks.

Visible light communication (VLC) has recently emerged as a promising solution for 6G. In order to fully exploit the extensive coverage capabilities of VLC, multi-user communication scenarios have garnered significant attention. However, existing techniques, such as light beam shaping using Spatial Light Modulators (SLMs), are limited by uncontrollable focusing planes and positioning errors, which result in signal attenuation and communication interruption. Digital holography, another approach, is hindered by factors like dependence on physical lenses and high crosstalk for multiplate imaging. To address these issues, this paper proposes a Random Multiplate Hologram (RMPH) technique that employs high-dimensional orthogonal random vectors to achieve depth-multiplexed multi-user holography. This method effectively mitigates crosstalk between holograms enhances system robustness, and increases signal acceptance probability (AP) when positioning errors are taken into account. The RMPH approach can accurately control the size and position of targets at different depths and achieve over 90% AP for users with varying reception ranges at distinct positions in simulations. Experimental results indicate that the RMPH method exhibits a power variation of only approximately 1 dB, compared to the multi-Fresnel lens (MF) method's power variation of up to 10.9 dB. Furthermore, the RMPH method demonstrates superior communication capabilities, achieving an 9.40 dB power improvement at the edge of the receiving range compared to other methods. Moreover, the RMPH method helps to reduce crosstalk among different depths. Compared to conventional method, the RMPH method shows an improvement of more than 4 dB in the received optical power when conducting forward and backward displacement. To the best of our knowledge, this is the first-time high-quality communications have been provided simultaneously to users at different 3D locations using digital holography. This outcome underscores the feasibility of future multi-user communication in the realm of visible light communication.

Fronthaul solutions based on optical wireless communication (OWC) have emerged as a promising technology for 6G communications. OWC systems offer advantages such as flexible deployment and low construction costs, but face significant challenges due to severe signal-signal beat interference (SSBI), nonlinear distortions and atmospheric turbulence. Efficiently optimizing these joint impairments is particularly crucial for large-scale deployment. Additionally, various mobile network architectures typically have distinct fronthaul requirements, and different applications may necessitate a certain level of isolation. In this paper, we propose a multi-band adaptive optimization autoencoder framework (MBAE) to achieve end-to-end (E2E) optimization for OWC-based fronthaul systems supporting multi-access architecture. The impairments from the multiple transceiver systems and the propagation link can be jointly optimized to achieve performance closest to the global optimum. Studies are conducted in a simulated channel and a real optical wireless communication system respectively to verify the feasibility of the proposed channel model and MBAE. Simulation results show that MBAE can obtain an average Q factor gain of 3 dB compared with the multi-band carrier-less amplitude and phase (CAP) modulation with neural network based post-equalization (NN) and decision feedback equalization (DFE). The experimental results show that MBAE can achieve a 1.9 dB received optical power (ROP) sensitivity gain compared with CAP using NN and DFE at a transmission rate of 36 Gbps under the 7% hard decision-forward error correction (HD-FEC) threshold of 3.8×10-3 utilizing a low-cost PD with 10G-class bandwidth and support a link budget of 44.7 dB. In the performance test under turbulent conditions, MBAE also exhibits superior average performance, achieving a mean Q factor gain of 1.41 dB compared with CAP using NN and DFE. The proposed method can be a promising solution to meet the high-speed and low-cost requirements of future optical wireless fronthaul systems.