A solar-blind multi-quantum well (MQW) structure wafer based on AlGaN materials is epitaxial growth by metal-organic chemical vapor deposition (MOCVD). The monolithically integrated photonic chips including light-emitting diodes (LEDs), waveguides, and photodetectors (PDs) are presented. The results of the finite-difference time-domain (FDTD) simulation confirm the strong light constraint of the waveguide designed with the triangular structure in the optical coupling region. Furthermore, in virtue of predominant ultraviolet transverse magnetic (TM) modes, the solar blind optical signal is more conducive to lateral transmission along the waveguide inside the integrated chip. The integrated PDs demonstrate sufficient photosensitivity to the optical signal from the integrated LEDs. When the LEDs are operated at 100 mA current, the photo-to-dark current ratio (PDCR) of the integrated PD is about seven orders of magnitude. The responsivity, specific detectivity, and external quantum efficiency of the integrated self-driven PD are 74.89 A/W, 4.22×1013 Jones, and 3.38×104%, respectively. The stable on-chip optical information transmission capability of the monolithically integrated photonic chips confirms the great potential for application in large-scale on-chip optical communication in the future.
Deep neural networks (DNN) are widely used in image recognition, image classification, and other fields. However, as the model size increases, the DNN hardware accelerators face the challenge of higher area overhead and energy consumption. In recent years, stochastic computing (SC) has been considered a way to realize deep neural networks and reduce hardware consumption. A probabilistic compensation algorithm is proposed to solve the accuracy problem of stochastic calculation, and a fully parallel neural network accelerator based on a deterministic method is designed. The software simulation results show that the accuracy of the probability compensation algorithm on the CIFAR-10 data set is 95.32%, which is 14.98% higher than that of the traditional SC algorithm. The accuracy of the deterministic algorithm on the CIFAR-10 dataset is 95.06%, which is 14.72% higher than that of the traditional SC algorithm. The results of Very Large Scale Integration Circuit (VLSI) hardware tests show that the normalized energy efficiency of the fully parallel neural network accelerator based on the deterministic method is improved by 31% compared with the circuit based on binary computing.
Wavelength selective switch (WSS) is the crucial component in the reconfigurable optical add/drop multiplexer (ROADM), which plays a pivotal role in the next-generation all-optical networks. We present a compact architecture of twin 1×40 liquid crystal on silicon (LCoS)-based WSS, which can be regarded as a 4f system in the wavelength direction and a 2f system in the switching direction. It is designed with theoretical analysis and simulation investigation. Polarization multiplexing is employed for two sources of twin WSS by polarization conversion before the common optical path. The WSS system attains a coupling efficacy exceeding 96% for 90% of the ports through simulation optimization. The 3 dB bandwidth can be achieved by more than 44 GHz at a 50 GHz grid for all 120 channels at all deflection ports. This work establishes a solid foundation for developing high-performance WSS with larger port counts.
With the advancement of photonic integration technology, ultra-low linewidth frequency-stabilized lasers have demonstrated significant potential in precision measurement, quantum communication, atomic clocks, etc. This review summarizes the latest developments in integrated photonics for achieving ultra-low linewidth lasers, particularly breakthroughs made by integrating Brillouin lasers. We discuss the design principles, manufacturing processes, performance characteristics, and potential value of these lasers in various applications.
GaN-based devices have developed significantly in recent years due to their promising applications and research potential. A major goal is to monolithically integrate various GaN-based components onto a single chip to create future optoelectronic systems with low power consumption. This miniaturized integration not only enhances multifunctional performance but also reduces material, processing, and packaging costs. In this study, we present an optoelectronic on-chip system fabricated using a top-down approach on a III-nitride-on-silicon wafer. The system includes a near-ultraviolet light source, a monitor, a 180° bent waveguide, an electro-absorption modulator, and a receiver, all integrated without the need for regrowth or post-growth doping. 35 Mbit/s optical data communication is demonstrated through light propagation within the system, confirming its potential for compact GaN-based optoelectronic solutions.
While considerable research has been conducted on the structural principles, fabrication techniques, and photoelectric properties of high-voltage light-emitting diodes (LEDs), their performance in light communication remains underexplored. A high-voltage series-connected LED or photodetector (HVS-LED/PD) based on the gallium nitride (GaN) integrated photoelectronic chip is presented in this paper. Multi-quantum wells (MQW) diodes with identical structures are integrated onto a single chip through wafer-scale micro-fabrication techniques and connected in series to construct the HVS-LED/PD. The advantages of the HVS-LED/PD in communication are explored by testing its performance as both a light transmitter and a PD. The series connection enhances the device's 3 dB bandwidth, allowing it to increase from 1.56 MHz to a minimum of 2.16 MHz when functioning as an LED, and from 47.42 kHz to at least 85.83 kHz when operating as a PD. The results demonstrate that the light communication performance of HVS-LED/PD is better than that of a single GaN MQW diode with bandwidth and transmission quantity, which enriches the research of GaN-based high-voltage devices.
Vision-based measurement technology benefits high-quality manufacturers through improved dimensional precision, enhanced geometric tolerance, and increased product yield. The monocular 3D structured light visual sensing method is popular for detecting online parts since it can reach micron-meter depth accuracy. However, the line-of-sight requirement of a single viewpoint vision system often fails when hiding occurs due to the object’s surface structure, such as edges, slopes, and holes. To address this issue, a multi-view 3D structured light vision system is proposed in this paper to achieve high accuracy, i.e., Z-direction repeatability, and reduce hiding probability during mechanical dimension measurement. The main contribution of this paper includes the use of industrial cameras with high resolution and high frame rates to achieve high-precision 3D reconstruction. Moreover, a multi-wavelength (heterodyne) phase expansion method is employed for high-precision phase calculation. By leveraging multiple industrial cameras, the system overcomes field of view occlusions, thereby broadening the 3D reconstruction field of view. Finally, the system achieves a Z-axis repetition accuracy of 0.48 μm.
Passive intermodulation (PIM) in communication systems is an unwanted interference caused by weak nonlinear current-voltage characteristics of radio frequency (RF) passive components. Characterization of PIM is important for both the study of PIM mechanisms and the location/suppression of PIM sources. PIM probes, made of open-ended coaxial transmission lines, have almost the same coupling strength to carriers and PIM products, and are usually used for near-field PIM characterization. Namely, it doesn’t have any filtering capability. Therefore, it cannot stop the carrier power from entering into PIM tester’s receiver, which may trigger active intermodulation of the receiver and degrade the PIM tester’s performance. To overcome this drawback, a passive filtering coaxial probe is proposed here. Compared with existing passive coaxial PIM probes, it has stronger coupling strength for PIM products than for carriers. Thus, the probe itself can block part of the carrier power entering into the PIM tester’s receiver. This advantage helps improve PIM tester’s overall performance. Both theoretical analysis and experiments are conducted for demonstration. The proposed probe brings more possibility to PIM characterization.
We propose a dynamic simultaneous localization and mapping technology for unsupervised motion removal (UMR-SLAM), which is a deep learning-based dynamic RGBD SLAM. It is the first time that a scheme combining scene flow and deep learning SLAM is proposed to improve the accuracy of SLAM in dynamic scenes, in response to the situation where dynamic objects cause pose changes. The entire process does not require explicit object segmentation as supervisory information. We also propose a loop detection scheme that combines optical flow and feature similarity in the backend optimization section of the SLAM system to improve the accuracy of loop detection. UMR-SLAM is rewritten based on the DROID-SLAM code architecture. Through experiments on different datasets, it has been proven that our scheme has higher pose accuracy in dynamic scenarios compared with the current advanced SLAM algorithm.
To enhance the video quality after encoding and decoding in video compression, a video quality enhancement framework is proposed based on local and non-local priors in this paper. Low-level features are first extracted through a single convolution layer and then processed by several conv-tran blocks (CTB) to extract high-level features, which are ultimately transformed into a residual image. The final reconstructed video frame is obtained by performing an element-wise addition of the residual image and the original lossy video frame. Experiments show that the proposed Conv-Tran Network (CTN) model effectively recovers the quality loss caused by Versatile Video Coding (VVC) and further improves VVC's performance.
As an important branch of federated learning, vertical federated learning (VFL) enables multiple institutions to train on the same user samples, bringing considerable industry benefits. However, VFL needs to exchange user features among multiple institutions, which raises concerns about privacy leakage. Moreover, existing multi-party VFL privacy-preserving schemes suffer from issues such as poor reliability and high communication overhead. To address these issues, we propose a privacy protection scheme for four institutional VFLs, named FVFL. A hierarchical framework is first introduced to support federated training among four institutions. We also design a verifiable replicated secret sharing (RSS) protocol 32-sharing and combine it with homomorphic encryption to ensure the reliability of FVFL while ensuring the privacy of features and intermediate results of the four institutions. Our theoretical analysis proves the reliability and security of the proposed FVFL. Extended experiments verify that the proposed scheme achieves excellent performance with a low communication overhead.
