A high-resolution ranging system with an extended measurement range is proposed and demonstrated for optical fibers, which breaks through the inherent coherence-length limitation of conventional schemes. In the proposed system, photonics-assisted signal processing is utilized to suppress optical phase noise and improve range resolution effectively. Meanwhile, a low-coherence optical source is adopted to suppress multi-echo interference among multiple optical paths. Experimental results show that the range resolution is immune to optical phase noise and remains stable over varying measurement ranges. For a fiber under test of approximately 10 km, a range resolution of 2.5 cm is realized, validating its potential for advanced metrology applications.
Optical see-through head-mounted displays (OST HMDs) suffer from significant degradation in image quality due to uncontrolled external light. To address this, we propose a perceptually optimized local dimming pattern based on intermediate virtual images. Unlike conventional simple patterns that directly mimic the virtual object shape, our gradient-based mask design leverages the amplitude profile of the intermediate image to reduce both color distortion and defocus blur. To evaluate effectiveness, we introduce a segmentation-based image similarity assessment using a perceptual image similarity metric (LPIPS). Experimental results show that, under equal blur conditions, the proposed pattern achieves improved color fidelity over simple patterns within a practical mask-to-eye distance range. This method offers a compact and computationally efficient solution for enhancing visual quality in OST HMDs, particularly under strong ambient lighting.
This work demonstrates the experimental realization of a highly stable and selectable single-longitudinal-mode (SLM) erbium-doped fiber (EDF) ring laser. By integrating a quintuple-ring cavity induced mode-filtering effect, dense longitudinal-mode oscillations are suppressed, yielding linewidths of 312.5 to 625 Hz and extend the wavelength sweep span from 1514.0 to 1565.0 nm. The flat power output is also obtained based on the designed ring architecture. Additionally, the optical signal-to-noise ratio (OSNR), output power, and instability of the EDF ring laser are evaluated and analyzed.
This letter proposes a traffic-adaptive dual-mode optical code-division multiplexing (OCDM) labeling scheme for optical packet switching (OPS). The scheme integrates complement-augmented on-off keying (CA-OOK) and two-code keying (TCK) within a unified spectral-amplitude coding (SAC) framework using arrayed-waveguide grating (AWG)-based codecs. TCK improves label detection reliability under light traffic, whereas CA-OOK expands the label set and alleviates capacity overflow under heavy traffic. A packet-loss probability (PLP) analysis identifies the traffic-dependent crossover between the two modes and guides mode selection. Numerical results show that the proposed adaptive scheme achieves a lower PLP than fixed-mode labeling schemes over a wide range of traffic conditions.
To extend the low-refractive-index (RI) detection capability of optical fiber-based surface plasmon resonance (SPR) sensors, we propose an SPR sensor based on a no-core fiber (NCF) externally coated with an Indium Tin Oxide (ITO) film, which achieves a substantial red-shift of the resonance peak—enabling detection over a low RI range of 1.00 to 1.34—while the simple structure of the NCF reduces fabrication complexity, and the location of the resonance wavelength within the telecommunications band allows the use of mature commercial equipment to lower manufacturing costs. Furthermore, our theoretical analysis of the influence of metal film material (ITO versus gold) and ITO thickness on sensor performance reveals that, for RI values below 1.35, the ITO film offers enhanced performance compared to the gold film, and that increasing the ITO thickness improves sensitivity at the expense of a reduced detection upper limit.
Using a cheap DFB laser with coherence length Lc =7 m and reading the field returning from a remote target by means of a Self-Mixing Interferometry (SMI) scheme, we are able to sense returns and measure distances up to 60 m thanks to the (1+C)2 factor of increase of Lc. due to optical feedback into the laser cavity. The simplified hardware and low cost of the SMI configuration makes it attractive as medium distances LiDAR for automotive applications.
We propose a dynamic joint biological encryption scheme based on a feature-driven 5-D cascaded chaotic system. By extracting statistical variance and distribution proportions from joint biological encoding rules derived from the 3-D Lorenz system, the scheme dynamically controls the parameters and initial states of a subsequent 2-D HSM map. The scheme achieves a vast key space of 10150. Experimental validation over a 25-km standard single-mode fiber (SSMF) demonstrates a 3.75 Gbps rate for 16QAM-OFDM signals, guaranteeing reliable data recovery for authorized users at receiver sensitivities above -14 dBm while strictly denying unauthorized access. Consequently, this scheme significantly bolsters stability and robustness, offering a highly secure physical-layer encryption scheme for advanced optical communications.
We experimentally assess the performances of default, high margin and high throughput LDPC options in ITU-T 50G-PON at 25 Gb/s. We also demonstrate that a 2.6 dB margin can be obtained with an even lower code-rate while ensuring, nonetheless, net throughput compatibility with legacy XGS-PON.
The switching dynamics of semi-insulating gallium nitride (GaN) photoconductive semiconductor switches (PCSSs) are found to be trap-limited. Under direct current bias up to 3 kV and laser energies from 0.1 to 3 mJ, the photocurrent increases nonlinearly with laser energy and saturates: a 30-fold increase in laser energy yields less than a 4-fold increase in photocurrent. This trap-limited behavior arises from deep-level trap filling, Auger recombination, and electric - field screening. The pulse width shows no monotonic dependence on bias voltage due to competition between carrier sweep-out and trap filling, but increases monotonically with laser energy as trap saturation prolongs the effective carrier lifetime. These results provide quantitative guidelines for optimizing GaN PCSS triggering conditions1.
A polarization-insensitive fiber optic parametric amplifier (FOPA) is demonstrated based on a polarization diverse bidirectional loop consisting of two equal-length highly nonlinear polarization maintaining fiber (PMF) sections and a 4-port polarization beam splitter (PBS). With equal pump power in the clockwise and counterclockwise directions, polarization sensitivity of OPA gain can be avoided without the need of polarization controlling in the loop. By 90° axis rotation between the two PMF sections the impact of polarization mode dispersion is minimized. Because the states of polarization of the counter propagating pumps in the loop are mutually orthogonal, stimulated Brillouin scattering (SBS) generated by the pump in each direction is not amplified by the parametric process of the pump in the opposite direction, which effectively eliminated the interference noise caused by pump/SBS mixing. System performance is investigated by inserting the FOPA in front of a dual-polarization 400Gb/s (16-QAM) coherent receiver as a preamplifier. Good agreement between measured and calculated BER versus received signal power indicated negligible polarization-dependent gain of the FOPA.
This paper proposes a design method for a passive fiber-optic current sensor (FOCS) capable of overcoming dynamic range limitations. By employing a dual-path Sagnac sensing structure, the system generates two delayed, orthogonal pulsed interference signals. Through least-squares ellipse fitting and correction to eliminate asymmetric optical errors, the linear intervals of the two signals are seamlessly stitched together. Theoretically, this approach breaks through inherent dynamic range constraints to achieve an ultra-large measurement span. Experimental results demonstrate that, although constrained by current laboratory conditions, the absolute measurement range of the proposed system reaches three times that of traditional open-loop passive Sagnac sensors, achieving an effective dynamic range of 60.92 dB. By completely discarding active components, the proposed scheme offers a simpler architecture and lower cost compared to mainstream closed-loop feedback FOCS, while simultaneously realizing a significantly broader dynamic range than conventional open-loop configurations.
This letter proposes a high-linearity, large-swing modulator driver design methodology for advanced CMOS processes in high-level pulse amplitude modulation (PAM) optical transmitters. Targeting a commercial thin-film lithium niobate (TFLN) Mach-Zehnder modulator (MZM) and an electro-absorption modulated laser (EML), the driver employs a current-mirror-based push-pull stacked output stage (CMPPS) and a dynamic current bleeder (DCB). The CMPPS overcomes breakdown voltage limits for broadband high-swing operation, while the DCB enables analog pre-distortion by restoring effective output current under large-signal conditions, improving swing and linearity with minimal power penalty. Fabricated in 28-nm standard CMOS, the prototype achieves 140-Gbps PAM-4 with a level separation mismatch ratio (RLM) of 0.94 and extinction ratio (ER) of 5.2 dB using the MZM and 112-Gbps PAM-4 with an RLM of 0.91 and ER of 6.7 dB using the EML.
We report an experimental demonstration that suitable polarization control of the feedback light, reflected by a fiber Bragg grating (FBG) back into the 974 nm pump diode, effectively suppresses the pump relative intensity noise (RIN), thereby improving the noise performance of a stretched pulse fiber laser. When adjusting the pump feedback polarization, the optical signal-to-noise ratio (OSNR) of pump increases by over 10 dB and its RIN decreases by ~15 dB. As a result, the corresponding generated mode-locked pulses exhibit a reduction of ~15 dB in low-frequency RIN, and the integrated timing jitter (100 Hz-100 kHz) decreases from 45.47 fs to 18.31 fs, while the pulse characteristics remain essentially unchanged. To the best of our knowledge, this represents the lowest low-frequency RIN achieved in a stretched pulse fiber laser operating at 1.5 μm. It provides a novel and practical method for noise reduction in ultrafast fiber lasers, gaining potential for high-precision applications.
Although deep learning-based image denoising achieves excellent performance, its massive parameters and computational complexity hinder deployment in low-power edge systems. To address these challenges, this paper proposes an Optoelectronic Attention Denoising Network (OEADNet) for color images. It consists of a single-layer multi-kernel parallel metasurface optical front-end for feature extraction and a lightweight Restormer-style electronic back-end for restoration. By combining a Restormer teacher network with spatial domain result distillation, it achieves end-to-end collaborative optimization. The model requires 0.364 M parameters and 21.72 G FLOPs, accounting for only 1.39% and 7.70% of the teacher network. Numerical experiments demonstrate stable restoration across various Gaussian noise levels. At a noise level of σ=50 for 8-bit images, OEADNet improves upon a same-scale pure electronic baseline by up to 1.41 dB, and surpasses denoising convolutional neural networks (DnCNNs) by 0.77 dB. Furthermore, it extends optical neural networks from grayscale scenes to color denoising, providing a design that balances performance and physical realizability for resource-constrained systems.
We propose a photonic topological optimization (TO) method based on Gaussian Basis Function (GBF) parameterization. By representing the design domain as a superposition of GBFs, this approach replaces the physically-meaningful density with a freely-growing field as the geometry descriptor that drives binarization. Combined with a soft-threshold projection into a full-domain gradient chain, this approach facilitates autonomous topological growth and accelerates optimization progress. The intrinsic low-pass filtering of GBF eliminates auxiliary spatial filtering needs, yielding fabrication-friendly structures with suppressed high-frequency spatial components. Validated on Y-branch splitters and waveguide crossings, numerical results demonstrate that our method suppresses high-frequency spatial components by nearly 50% and reduces iteration counts by 70% with comparable performance. This method offers a versatile and efficient framework for the inverse design of high-performance photonic integrated devices.
In this Letter, we propose and demonstrate a dual-mode polarization index modulation (DM-PIM) scheme for polarization division multiplexing optical wireless communication (PDM-OWC) systems. By performing dual-mode index modulation in the polarization domain, DM-PIM can transmit both index and constellation bits to enhance the spectral efficiency of PDM-OWC systems. As constellation design plays a vital role in DM-PIM, we further propose and optimize two dual-mode constellations based on pulse amplitude modulation (PAM) format. Since PDM-OWC systems are highly vulnerable to polarization misalignment, we also evaluate the impact of polarization misalignment on the performance of PDM-OWC systems employing different transmission schemes. Experimental results show that DM-PIM with optimized dual-mode PAM constellation exhibits superior bit error rate (BER) performance and strong robustness against polarization misalignment in PDM-OWC systems.
To mitigate the impact of atmospheric turbulence in long-distance optical camera communication (OCC) systems, a Manchester coding-based demodulation method is proposed. A rolling-shutter-based OCC transmission experiment is conducted over a distance of 2.8 km. At the transmitter, the data are encoded with Manchester code, while at the receiver, the inherent transition characteristics of Manchester codes are exploited for demodulation, thereby improving demodulation stability under atmospheric turbulence conditions. The proposed demodulation method is simple without decision threshold, and exhibits strong interference robustness. The experimental results show that the bit error rate (BER) of the proposed method is maintained below the hard-decision forward error correction (HD-FEC) threshold under both weak and strong turbulence conditions, and data transmissions of 4.5 kb/s and 4.3 kb/s are achieved, respectively, with the data transmission rate being increased by 21% and 30% approximately, compared with the polynomial fitting threshold demodulation method.
Accurate quality of transmission (QoT) estimation, particularly generalized signal-to-noise ratio (GSNR) prediction, in newly deployed C+L-band optical networks is hindered by data scarcity and incomplete physical link information, leading to a cold-start problem for conventional deep learning methods. To address this issue, we propose PAL-MISA, a data-efficient framework combining Parameterized Active Learning (PAL) and Meta-Initialized Sparse Adaptation (MISA). PAL learns to select high-value measurements, while MISA provides a transferable initialization for rapid sparse adaptation to unseen physical environments. Validated on three representative network topologies, PAL-MISA accelerates convergence and reduces performance fluctuations compared with conventional training from scratch in the target domain. To achieve the same prediction error, PAL-MISA reduces the required measurement data volume by 80% and ultimately achieves a minimum mean absolute error (MAE) of 0.04 dB, offering a robust solution for digital-twin deployment in uncharacterized optical networks.
We propose a high-performance correlation OTDR scheme based on complementary Golay codes for long-reach PON monitoring. By leveraging unipolar-to-bipolar mapping and orthogonal sidelobe cancellation, the scheme successfully overcomes the PSL saturation limits that constrain conventional single-sequence coding. Separating itself from single-sequence correlation which prematurely plateaus due to structural code noise, the proposed system effectively eliminates algorithmic sidelobes. Experimentally evaluating a 20-km feeder fiber with a 1:16 passive splitter, the system achieves a sub-meter spatial resolution of 0.6 m and a sensitivity of -30 dBm. With only 33 averages, the PSL avoids any saturation plateau and reaches 30.24 dB. Furthermore, the system delivers a 31.27-dB SNR and a 7.21-dB enhanced DR specifically for weak reflection peaks (in 5-log), demonstrating robust fault characterization for dense optical distribution networks.
