Matrix-vector multiplication (MVM) is the basic operation for photonic tensor core, which lies at the core of neural network inference and training, forming the dominant computational primitive underlying every layer of modern deep architectures. Although optical neural networks promise massive acceleration of these operations through parallel electromagnetic propagation, most existing implementations suffer from limited reconfigurability, off-chip footprint, or complex integration requirements. In this work, we introduce an on-chip computing element specifically tailored for neural-network MVM, constructed from a cascaded arrangement of metasurfaces. The metasurfaces incorporate vanadium dioxide and photoactivated GaAs, whose distinct switching mechanisms enable each layer to encode matrix weights and input vector elements through external electrical, thermal, and optical stimuli. As terahertz waves sequentially traverse the cascade, the system physically accumulates weighted contributions, yielding the equivalent multiplication between a 2 & times; 2 matrix and a 2 & times; 1 input vector-executed at wave speed. Finite integration time-domain simulations validate all logical combinations of matrix elements and inputs and confirm the correct arithmetic mapping for two independent computing channels. Moreover, it supports general weight programmability and maintains stable responses with respect to incident and polarization angles. By embedding the most fundamental neural-network operation directly into a planar metasurface stack, this work provides a compact pathway toward integrated, low-power optical accelerators and lays the groundwork for future metasurface-based deep learning hardware.

High-order modulation format optical signals are widely deployed in modern optical networks but remain insufficiently protected at the physical layer, leading to significant security vulnerabilities. This work proposes an all-optical secure transmission scheme for 16QAM signals based on phase coding and dispersive stretching. The scheme employs symmetric cascaded in-phase and quadrature Mach-Zehnder modulators (IQMZMs) to realize phase encoding and decoding, providing a simple structure with good suitability for photonic integration. The universality of constant-envelope key signals within the proposed architecture is investigated, providing references for key-format and device selection in practical applications and important guidance for future research on all-optical digital encryption. To the best of our knowledge, this is the first demonstration of all-optical digital encryption and decryption of 16QAM signals using high-order modulated key signals. Compared with existing schemes of the same category, the key space increases by a factor of $2<^>{n}$, where $n$ denotes the number of symbols. Furthermore, the key signal undergoes dynamic temporal broadening through dispersive processing, enabling steganographic transmission. Simulation results at a bit rate of 100 Gb/s demonstrate significant security enhancement with high stability and reliability. Overall, the proposed scheme provides an effective approach to enhancing physical-layer security in high-order modulation optical transmission systems.

Conventional path protection algorithms in all-optical data center networks (DCNs) need a relatively long calculation time but with low efficiency. In this article, we propose a preset path recovery algorithm by combining machine learning and priority execution method. The former is to predict areas of frequent optical path failures in data centers, and the latter is to ensure accurate protection in those areas while reducing computation time and improving performance. Firstly, we construct a dataset using rules composed of topological edge relationships, failure time, failure frequency, packet loss rate, latency and service type. Then, based on the above dataset, SVM (support vector machine) algorithm is used to predict frequent failure areas. Finally, based on different prediction frequencies, priority construction is carried out to cluster the entire topology, and p-cycle (preconfigured protection cycle) is applied to each topology. Experiments show that the combination of SVM prediction and classification protection can significantly reduce the protection range of potential failure areas to reduce computation time while ensuring high accuracy, as compared with traditional optical path protection algorithms.

With the explosion of data traffic triggered by 5G/6G and Generative artificial intelligence, coherent optical communication is moving towards higher baud rates and more complex modulation formats. This leads to a significant increase in the computational complexity and power consumption of digital signal processing (DSP) at the transmitter and receiver ends, especially in the chromatic dispersion(CD) Compensation and low roll-off shaping filter modules. We propose a joint shaping filtering and CD compensation (JFS-CD) algorithm. This algorithm moves the CD compensation to the transmitter side and utilizes the characteristics of discrete fourier transform and the spectral features of shaping filtering for integrated processing. Aiming at the high peak-to-average power ratio (PAPR) problem caused by chromatic dispersion pre-compensation, we propose a low-complexity square boundary clipping algorithm(SBC). Simulation results show that, under the premise of maintaining unchanged performance, JFS-CD can reduce the real multiplication complexity by about 46

The surge in data traffic catalyzed by generative artificial intelligence(AI) and 5G/6G services has placed unprecedented demand on data center interconnects (DCI). To meet these requirements, coherent optical systems for DCI applications are rapidly scaling toward higher baud rates and advanced modulation formats, necessitating more efficient digital signal processing (DSP) to balance performance and complexity, especially in chromatic dispersion (CD) compensation. We investigate a joint pulse shaping and CD pre-compensation (JPS-CD) architecture for coherent DCI systems. By exploiting the DFT scaling property and the spectral sparsity of low-roll-off raised-cosine pulse shaping, JPS-CD reduces unnecessary frequency-domain multiplications outside the occupied signal bandwidth. We further show that pure CD pre-compensation increases the waveform PAPR and reduces the transmitter SNR after peak normalization; therefore, a low-complexity square boundary clipping (SBC) algorithm is introduced to mitigate this limitation. Simulation and experimental results show that JPS-CD reduces the real multiplication complexity by about 41% while maintaining comparable performance.

Objective Optical switches play a crucial role in routing optical signals between different paths. Currently, most mode-division multiplexing (MDM)-based optical switches rely on thermo-optic or electro-optic effects. This limitation leads to device dimensions typically reaching several hundred micrometers, constraining the scalability of optical switches. Additionally, the need for a continuous power supply to maintain the switching state results in high power consumption. To address these challenges, this paper proposes a high-refractive-index-contrast Sb2Se3-Si hybrid waveguide and demonstrates a non-volatile, scalable 1 & times;3 mode (de)-multiplexer based on MDM and phase-change material (PCM) technologies. The proposed design aims to expand link transmission capacity, reduce structural dimensions, lower static power consumption, and enable flexible routing of signals across different modes. Methods This paper first designs a high-refractive-index-contrast Sb2Se3-Si hybrid waveguide by etching the silicon waveguide to a specific thickness and depositing Sb2Se3 into the etched grooves. The 1 & times;3 mode demultiplexer consists of a central multi-mode bus waveguide flanked by single-mode waveguides above and below it. Demultiplexing is achieved when higher-order mode signals from the multi-mode waveguide are converted into fundamental-mode signals in the single-mode waveguides. Conversely, multiplexing occurs when fundamental-mode signals from the single-mode waveguides are converted back into higher-order modes in the multi-mode bus waveguide. The 1 & times;2 multimode optical switch integrates the six scalable 1 & times;3 mode (de) multiplexers. At each stage, the number of supported modes in the multi-mode bus waveguide decreases sequentially. Adiabatic tapered waveguides connect adjacent stages to match waveguides with differing cross-sectional parameters, minimizing energy loss. Results and Discussions The proposed method employs ANSYS Lumerical software for simulation and performance verification. First, we compared the effective refractive indices of Sb2Se3-Si hybrid waveguides under four distinct parameter configurations. To minimize device size, a high-refractive-index-contrast hybrid waveguide was implemented in the coupling regions of the 1 & times;3 mode (de) multiplexer. Furthermore, to enhance link capacity, a 1 & times;2 multimode optical switch was constructed by cascading stages of the mode (de) multiplexer using adiabatic tapered waveguides, ensuring efficient mode conversion. Simulation results demonstrate that, compared to the 20/220 nm Sb2Se3/Si configuration, the 100/120 nm Sb2Se3/Si configuration achieves transmission improvements of 34 degrees o, 7.6 degrees o, 19.7 degrees o, 30.4 degrees o, 27 degrees o, and 20.5 degrees o for the six output ports of the 1 & times;3 demultiplexer, respectively. Across the entire C-band, the insertion loss (IL) remains below 0.5 dB for all modes, with crosstalk (CT) suppressed to below-9.6 dB. Additionally, the total IL of the multimode switch structure is below 0.6 dB. Conclusions This paper presents a high-refractive-index-contrast Sb2Se3-Si hybrid waveguide and designs non-volatile, scalable 1 & times;3 mode (de)multiplexers. By cascading these (de)multiplexer stages, we demonstrate a 1 & times;2 multimode optical switch capable of six-mode operation. The proposed hybrid waveguide design achieves significant reductions in both multimode waveguide width and coupling length. Furthermore, the scalable 1 & times;3 (de)multiplexer architecture substantially decreases static power consumption while increasing link transmission capacity. Under our optimized design parameters, the 1 & times;2 multimode switch exhibits low insertion loss (<0.6 dB) across the entire C-band and enables flexible routing of different modes to either output port. The incorporation of adiabatic tapered waveguides for inter-stage connections further minimizes energy loss, resulting in superior overall performance.

To meet the high demand for real-time processing and low nuisance alarm rates in distributed fiber vibration sensing systems based on phase-sensitive optical time domain reflectometry (phi-OTDR), a computationally efficient signal recognition method based on successive variational mode decomposition (SVMD) and a binary-tree-structured SVM classifier is proposed. Addressing the limitations of existing approaches (e.g., >8% false alarm rates caused by background noise interference), the proposed method performs adaptive-threshold SVMD to suppress non-vibration noise, extracts physically interpretable features from intrinsic mode functions (IMFs), and employs an optimized binary tree SVM classifier to reduce computational complexity. The experimental verification conducts on a 10.05 km field-deployed fiber demonstrates 97 % macro-averaged recognition accuracy for knocking (97.51 %), digging (95.93 %), shaking (96.43 %), watering (95.80 %), and noise events (99.50 %). Compared with the OAO-SVM algorithm, the recognition delay has been reduced by 64 % (8.5 ms per sample), and the false alarm rate in the on-site test is lower than 2.5 %. This method provides a deployable solution for distributed fiber alarm systems with limited training samples.

This paper proposes a physics-informed convolutional network (PICN) scheme to detect bending eavesdropping attacks in dual-polarization coherent optical communication systems. We present a theoretical model for optical signal transmission under bending eavesdropping, analyzing the impact of bending eavesdropping on fiber physical characteristics such as dispersion and nonlinear effect. These physical characteristics are embedded into a convolutional neural network (CNN) to construct PICN, which automatically captures subtle variations of the signal features under bending eavesdropping. To validate the effectiveness of the scheme, we first develop an eavesdropping experimental platform in an 80-km 168 Gbps dual-polarization quadrature phase shift keying (QPSK) coherent optical communication system. Polarization data are then collected under normal transmission, 10.8 mm and 15 mm bending radius. Finally, the detection performance of four classifiers including PICN, random forest (RF), support vector machine (SVM), and K-nearest neighbor (KNN) are evaluated at single and mixed bending radii. Experimental results demonstrate that PICN achieves detection accuracies of 100%, 98.53%, and 99.02% under 10.8 mm, 15 mm, and mixed bending radii, respectively. Our work provides novel theoretical foundations and innovative perspectives for bending eavesdropping detection in optical fiber communication systems.

Analyzing the optical time domain reflectometer (OTDR) test curves to determine the type of optical fiber events is an important basis for ensuring the operation quality of communication lines. Aiming at the problems of complex processes, long time-consumption, and low accuracy of traditional event identification methods, an automatic identification method of optical fiber events based on fast dynamic time warping (Fast-DTW) is theoretically proposed and experimentally verified. This work uses the wavelet transform (WT) and cell-averaging constant false alarm rate (CA-CFAR) algorithm to locate the events data collected by OTDR. The event signals are extracted according to the locations, and the Fast-DTW algorithm is utilized to judge the event type based on the calculated distorted path distance between the extracted event signals and the event templates. The proposed algorithm solves the problem of requiring a large number of templates due to the traditional correlation matching method requiring equal signal length and template length, greatly reducing the system’s resource occupation. It can achieve accurate positioning of event positions and quickly identify event types.

Optical temperature sensors with stretchability play a crucial role in the development of continuous, stable, and non-invasive wearable health monitoring systems. However, designing efficient and stretchable optical temperature sensors presents significant challenges. This study proposes a wearable optical temperature sensor based on flexible optical fibers, leveraging the temperature-sensitive properties of down-converted luminescent particles (ZnS:Mn). The flexible optical fibers, made from highly elastic polymers, can withstand tensile deformations of up to 250% and feature a core-cladding structure that effectively confines optical transmission. ZnS:Mn emits dual-wavelength light with distinct temperature dependencies, enabling stable temperature sensing through the intensity ratio. Experimental results demonstrate the sensor's exceptional temperature sensitivity, stability, and repeatability within the range of 7-80 degrees C. Notably, the sensor also exhibits the ability to rapidly detect body temperature and recognize respiratory patterns. This work offers a promising solution for the development of advanced, personalized medical and wearable health monitoring devices.

This review provides a comprehensive survey of the most recent developments in metasurfaces for applications in domains including wireless-optical switching and communications. In particular, we focus on discussion of multi-parameter optical field regulation and potential applications in system performance enhancement. By designing nanostructured arrays with specific geometries, metasurfaces can be used to effectively manipulate parameters including phase, amplitude, and polarization, thereby enabling the switching, transmission, testing, analysis, and processing of optical signals. Notably, the introduction of phase-change materials offers a novel approach that allows metasurfaces to achieve more flexible wireless-optical switching at higher speeds. In wireless-optical communication systems, multiplexing of the different degrees of freedom of the light beams can improve the data transmission capacity and rate significantly. Finally, we present our own metasurface design with its unique passive parallel beam splitting capacity, and we demonstrate the superiority of this design in applications including wireless-optical inter-rack connections in data centers and industrial inspection based on optical cross-connectors.

A graphene-metasurface half-adder based on plasmon-induced transparency is proposed here for the first time. The device was simulated and optimized using the finite integration time domain (FITD) method, and its performance was systematically evaluated. The half-adder achieves a maximum modulation depth (MD) of 85.57%, a maximum extinction ratio (ER) of 8.41 dB, and a minimum insertion loss (IL) of 0.19 dB. By encoding the Fermi level of graphene and the polarization angle of the incident terahertz wave as the two logic inputs, the half-adder produces the corresponding output states at 2.608 THz and 2.956 THz. The proposed metasurface design is structurally simple and performs robustly in realizing half-adder functionality, offering a promising new strategy for developing optical logic devices.

The development of flexible optical strain sensors is of great significance for accurately capturing complex human movements and achieving high-precision activity recognition. This work proposes a wearable sensor based on a Z-shaped optical micro/nano fibers (MNF), which combines a sandwich structure consisting of a central MNF sensing layer and two side polydimethylsiloxane encapsulation layers to achieve high sensitivity, ultrathin flexibility, and biocompatibility. With the MNF bending loss characteristic, the sensor converts external pressure or deformation into changes in light transmittance, with fast response capability (<100 ms) and repeatability. Further experimental results show that the sensor has high linearity in the pressure range of 0.1-1 N and the angle range of 0-17 degrees, with sensitivities of -65.14 %/N and -1.04 %/degrees, respectively. Additionally, the sensor attached to the skin surface successfully monitors gestures, pronunciation, swallowing, and breathing. Machine learning algorithms have been utilized to further enhance the accuracy of gesture recognition, enabling the encoding of gestures signals and English letters based on Morse code protocol for information transmission. This design provides a low-cost, anti-electromagnetic interference solution for high-precision physiological monitoring and human-machine interaction.

All-optical encryption provides high-speed and low-latency security at the optical layer, but its adaptability to multi-wavelength transmission and multiple modulation formats remains a major challenge. This paper proposes a novel phase mutual-encoding architecture for the all-optical encryption of M-PSK signals. The proposed system significantly reduces system complexity in multi-wavelength scenarios, offers reconfigurability, and demonstrates promising potential for integration. Multiple cascaded IQ Mach-Zehnder modulators (IQMZMs) are used at the transmitter for cross-encoding of two phase signals, while four-wave mixing (FWM) enables decryption via a high-nonlinear fiber (HNLF) at the receiver. Simulation results demonstrate that the system can enhance physical-layer security for both QPSK and 8PSK modulation formats through parameter adjustment alone, without modifying the overall system architecture. Stable transmission performance is maintained at a single-wavelength symbol rate of up to 80 Gbaud. Overall, the proposed scheme provides an effective solution for flexible physical-layer security enhancement in wavelength-division multiplexing (WDM) systems.

To enhance the throughput of state-of-the-art C+L-band optical networks, three significant factors are considered: spectral efficiency (SE) enhancement, margin reduction, and resource provisioning techniques that coordinate the benefits of these two factors. For margin reduction, an accurate physical-layer impairments (PLIs) model, namely the C+L-band enhanced component-wise Gaussian noise (C+L-ECWGN) model, is proposed to account for the modulation correction terms on PLIs and the impact of spectral shaping in C+L-band transmission, thereby providing a solid foundation for the application of probabilistic constellation shaping (PCS) in cross-layer network provisioning. The proposed model maintains a similar computational complexity to the traditional enhanced Gaussian noise (EGN) model, which neglects the actual spectral shape of demands while considering them as rectangular. The proposed model can reduce overestimation of modulation correction terms by up to 39.8% under typical roll-off conditions, with a maximum reduction in nonlinear interference (NLI) estimation error of 26.2% across various modulation formats compared with state-of-the-art models. Finally, by integrating PCS technology and the benefit of an accurate PLI estimate, the PLI-aware PCS cross-layer optimization (PAPCO) algorithm is developed to achieve continuous and adaptive modulation-to-SE mapping. Simulation results indicate that, compared with a benchmark scheme based on conventional EGN modeling with rectangular spectra and uniform QAM modulation, the proposed PAPCO-ECWGN algorithm reduces the normalized signal-to-noise ratio (SNR) margin to near zero, thus improving network throughput by 41.3% and average SE by 35.2 %.

In this paper, a low-complexity scheme for dual-polarization 16 quadrature amplitude modulation (DP-16QAM) transceiver IQ imbalance compensation is proposed, which reduces the effects of in-phase (I) and quadrature (Q) imbalance. This scheme enables the use of existing clock recovery loops and equalizers to probe the transceiver IQ skew. For transceiver IQ skew estimation, the receiver(RX) IQ skew is estimated by a clock recovery algorithm based on Gardner's timing error detection(GSMA), and the transmitter(TX) IQ skew is estimated by finding the value that yields the lowest equalizer error(Scanning delay algorithm, SDA). To address transmitter IQ amplitude/phase imbalance, we propose a low-complexity MIMO equalizer. It comprises a butterfly CV-MIMO and a non-butterfly two-layer real-valued MIMO based on a multimodulus algorithm (TMMA-RV-MIMO). For convenience, the new equalizer is called C-R AEQ. A 100 km transmission simulation and experiment with 36 Gbaud DP-16QAM signals showed that, with the TX/RX IQ skew estimation, the estimation error is less than 0.8/0.25 ps. Under IQ imbalance conditions, the C-R AEQ provides a 0.5 dB Q-factor gain compared with the 4×4 MIMO equalizer. In the absence of IQ imbalance, the C-R AEQ incurs a 0.3 dB Q penalty compared with the 4×4 MIMO equalizer. Furthermore, the C-R AEQ achieves a 47× reduction in real multiplications compared with the 4×4 MIMO equalizer.

Optical fiber communication networks are the backbone of information transmission, carrying over 90% of the world's data. However, optical fibers are susceptible to fiber-bending eavesdropping attacks, which can result in massive data leaks. The detection of fiber-bending eavesdropping attacks has become one of the effective strategies to safeguard data security. In this paper, we propose a detection scheme based on a convolutional neural network (CNN) for detecting fiber-bending eavesdropping attacks in dual-polarization coherent optical communication systems. To validate the feasibility and effectiveness of the proposed scheme, we first establish the relationship between the fiber-bending radius and the bending power loss coefficient. Based on this relationship, a 100 km, 25 GBaud dual-polarization 16 quadrature amplitude modulation (16QAM) coherent optical communication system is constructed. This system collects both normal polarization data and bending eavesdropping polarization data. Subsequently, the performance of the CNN algorithm in detecting fiber-bending eavesdropping attacks is evaluated. The results show that under bending radii of 10.8 mm, 12.1 mm, 15 mm, and mixed bending radii, the eavesdropping detection accuracy rates of the CNN reach 99.8%, 99.67%, 99.22%, and 96.15%, respectively.

The photonic firewall is a device that detects and localizes risky information at the optical layer to enhance the security of optical networks. The most significant and challenging component of the photonic firewall is the all-optical pattern matching system, which determines its overall performance. In response to the explosion of data volume, it is crucial to develop reconfigurable all-optical pattern matching systems capable of handling signals in arbitrary high-order modulation formats. All-optical pattern matching systems based on optical real-time Fourier transform(RTFT) and conjugate multiplication are proposed for phase-shift keying (PSK) optical signals. Numerical simulations demonstrate that the system is capable of matching 24-bit target sequences within 528-bit source sequences, which are signals in either QPSK or 8PSK modulation formats, at a transmission rate of 260 GBaud.