The You Only Look Once (YOLO) object detection network has garnered widespread adoption in various industries, owing to its superior inference speed and robust detection capabilities. This model has proven invaluable in automating production processes such as material processing, machining, and quality inspection. However, as market competition intensifies, there is a constant demand for higher detection speed and accuracy. Current FPGA accelerators based on 8-bit quantization have struggled to meet these increasingly stringent performance requirements. In response, we present a novel 4-bit quantization-based neural network accelerator for the YOLOv5 model, designed to enhance real-time processing capabilities while maintaining high detection accuracy. To achieve effective model compression, we introduce an optimized quantization scheme that reduces the bit-width of the entire YOLO network—including the first layer—to 4 bits, with only a 1.5% degradation in mean Average Precision (mAP). For the hardware implementation, we propose a unified Digital Signal Processor (DSP) packing scheme, coupled with a novel parity adder tree architecture that accommodates the proposed quantization strategies. This approach efficiently reduces on-chip DSP utilization by 50%, offering a significant improvement in performance and resource efficiency. Experimental results show that the industrial object detection system based on the proposed FPGA accelerator achieves a throughput of 808.6 GOPS and an efficiency of 0.49 GOPS/DSP for YOLOv5s on the ZCU102 board, which is 29% higher than a commercial FPGA accelerator design (Xilinx’s Vitis AI).

On-chip electrostatic discharge (ESD) protection poses a challenge in the chip fabrication process. In this study, a novel electric fuse (E-fuse) device featuring a simple structure of Ni metal on a SiO2 dielectric for ESD protection was proposed, and the physical mechanism of its operation was investigated in detail. Experimental evaluations, utilizing transmission line pulse (TLP) testing and fusing performance analyses, reveal that the E-fuse, constructed with a Ni metal layer measuring 5 μm in width, 100 μm in length, and 5 nm in thickness, achieved a significant ESD protection voltage of 251 V (VHBM) and demonstrates low-voltage fusing at a bias voltage of 7 V. Compared to traditional ESD protection devices, the E-fuse boasts a smaller size and removability. To assess fusing performance, devices of varying sizes were tested using a fusing lifetime model. This study supports both theoretical and empirical evidence, enabling the adoption of cost-effective, straightforward E-fuse devices for ESD protection.

Two-hierarchical nanostructures, characterized by two distinct configurations along the height direction, exhibit immense potential for applications in various fields due to their significantly enhanced controllable degree compared to single-order structures. However, due to the limitations imposed by planar technology, the realization of two-hierarchical nanostructures encounters huge challenges. In this work, we developed a one-step etching method based on inductively coupled plasma reactive ion etching for two-hierarchical nanostructures. Thanks to the shrinking effect of the Cr mask and the generation of a passivation layer during etching, the target materials experienced two different states from vertical etching to shrink etching. Consequently, the achieved two-hierarchical nanostructure configuration features a cross-section of an upper triangle and a lower rectangle, showing higher controllable degrees compared to the one-order ones. Both the mask pattern and etching parameters play crucial roles, by which two-hierarchical structures with diversiform shapes can be constructed controllably. This method for two-hierarchical nanostructures offers advantages including excellent control over structural properties, high processing efficiency, uniformity across large areas, and universality in materials. This developed strategy not only presents a simple and rapid nanofabrication platform for realizing optoelectronic devices, but also provides innovative ideas for designing the next generation of high-performance devices.

Semiconducting metal oxides are widely used for solar cells, photo-catalysis, bio-active materials and gas sensors. Besides the material properties of the semiconductor being used, the specific surface topology of the sensors determines device performance. This study presents different approaches for increasing the sensing area of semiconducting metal oxide gas sensors. Micro- and nanopatterned laser-induced periodic surface structures (LIPSSs) are generated on silicon, Si/SiO2 and glass substrates. The surface morphologies of the fabricated samples are examined by FE SEM. We selected the nanostructuring and characterization of nanostructured source Ni/Au and Ti/Au films prepared on glass using laser ablation as the most suitable of the investigated approaches. Surface structures produced on glass by backside ablation provide 100 nm features with a high surface area; they are also transparent and have high resistivity. The value of the hydrogen sensitivity in the range concentrations from 100 to 500 ppm was recorded using transmittance measurements to be twice as great for the nanostructured target TiO2/Au as compared to the NiO/Au. It was found that such transparent materials present additional possibilities for producing optical gas sensors.

Currently, there are various types of microscales and the conventional line detection system usually has only one detection method, which is difficult to adapt to the diverse calibration needs of microscales. This article investigates the high-precision measurement method of a microscale based on optoelectronics and the image integration method to solve the diversified calibration needs of microscales. The automatic measurement and processing system integrates two methods: the photoelectric signal measurement method and the photoelectric image measurement method. This article studies the smooth motion method based on ordinary linear guides, investigates the method of reducing the cosine error of a small-range interference length measurement, proposes an image-based line positioning method, and studies the edge and center recognition algorithms of the line. According to the experimental data, the system’s measurement accuracy was analyzed using the photoelectric signal measurement method to measure the 1 mm microscale, the maximum difference from the reference value was 0.105 μm, the standard uncertainty was 0.068 μm, and the absolute value of normalized error was less than 1. The accuracy of the image measurement method to measure the 1 mm microscale was consistent with that of the photoelectric signal method. The results show good consistency in the measurement results between the two methods of the integrated measurement system. The photoelectric signal method has the technical characteristics of high measurement efficiency and high accuracy, while the pixel-based measurement of the image method has two-dimensional measurement characteristics, which can realize measurements that cannot be realized by the photoelectric signal method; therefore, the measurement system of optoelectronics and image integration is characterized by high precision and a wide range of measurement adaptations.

The paper proposes a novel automatic control system for the attitude of the mini-satellites equipped with an actuator having N = 2 DGMSCMGs (Double Gimbal Magnetically Suspended Control Moment Gyros) in parallel and orthogonal configuration, as well as a DGMSCMG-type sensor for the measurement of the satellite absolute angular rate. The proportional-derivative controller, designed based on the Lyapunov-functions theory, elaborates the control law according to which the angular rates applied to the servo systems for the actuation of the DGMSCMGs gyroscopic gimbals are computed. The gimbal’s angular rates create gyroscopic couples acting on the satellite in order to control its attitude with respect to the local orbital frame. The new proposed control architecture was software implemented and validated, and the analysis of the obtained results proved the cancellation of the convergence errors and excellent angular rate precision.

In this paper, a novel AlGaN/GaN HEMT structure with a P-GaN buried layer in the buffer layer and a locally doped barrier layer under the gate (PN-HEMT) is proposed to enhance its resistance to single event transient (SET) effects while also overcoming the degradation of other characteristics. The device operation mechanism and characteristics are investigated by TCAD simulation. The results show that the peak electric field and impact ionization at the gate edges are reduced in the PN-HEMT due to the introduced P-GaN buried layer in the buffer layer. This leads to a decrease in the peak drain current (Ipeak) induced by the SET effect and an improvement in the breakdown voltage (BV). Additionally, the locally doped barrier layer provides extra electrons to the channel, resulting in higher saturated drain current (ID,sat) and maximum transconductance (gmax). The Ipeak of the PN-HEMT (1.37 A/mm) is 71.8% lower than that of the conventional AlGaN/GaN HEMT (C-HEMT) (4.85 A/mm) at 0.6 pC/µm. Simultaneously, ID,sat and BV are increased by 21.2% and 63.9%, respectively. Therefore, the PN-HEMT enhances the hardened SET effect of the device without sacrificing other key characteristics of the AlGaN/GaN HEMT.

In the framework of fully vertical GaN-on-Silicon device technology development, we report on the optimization of non-alloyed ohmic contacts on the N-polar n+-doped GaN face backside layer. This evaluation is made possible by using patterned TLMs (Transmission Line Model) through direct laser writing lithography after locally removing the substrate and buffer layers in order to access the n+-doped backside layer. As deposited non-alloyed metal stack on top of N-polar orientation GaN layer after buffer layers removal results in poor ohmic contact quality. To significantly reduce the related specific contact resistance, an HCl treatment is applied prior to metallization under various time and temperature conditions. A 3 min HCl treatment at 70 °C is found to be the optimum condition to achieve thermally stable high ohmic contact quality. To further understand the impact of the wet treatment, SEM (Scanning Electron Microscopy) and XPS (X-ray Photoelectron Spectroscopy) analyses were performed. XPS revealed a decrease in Ga-O concentration after applying the treatment, reflecting the higher oxidation susceptibility of the N-polar face compared to the Ga-polar face, which was used as a reference. SEM images of the treated samples show the formation of pyramids on the N-face after HCl treatment, suggesting specific wet etching planes of the GaN crystal from the N-face. The size of the pyramids is time-dependent; thus, increasing the treatment duration results in larger pyramids, which explains the degradation of ohmic contact quality after prolonged high-temperature HCl treatment.

Magnesium alloy ZK60 shows great promise as a medical metal material, but its corrosion resistance in the body is inadequate. Hydroxyapatite (HA), the primary inorganic component of human and animal bones, can form chemical bonds with body tissues at the interface, promoting the deposition of phosphorus products and creating a dense calcium and phosphorus layer. To enhance the properties of ZK60, HA was added to create HA/ZK60 composite materials. These composites, fabricated using the advanced technique of LPBF, demonstrated superior corrosion resistance and enhanced bone inductive capabilities compared to pristine ZK60. Notably, the incorporation of 3 wt% led to a significant reduction in bulk porosity, achieving a value of 0.8%. The Ecorr value increased from −1.38 V to −1.32 V, while the minimum Icorr value recorded at 33.9 μA·cm−2. Nano-HA achieved the lowest volumetric porosity and optimal corrosion resistance. Additionally, these composites significantly promoted osteogenic differentiation in bone marrow stromal cells (BMSCs), as evidenced by increased alkaline phosphatase (ALP) activity and robust calcium nodule formation, highlighting their excellent biocompatibility and osteo-inductive potential. However, when increasing the HA content to 6 wt%, the bulk porosity rose significantly to 3.3%. The Ecorr value was −1.3 V, with the Icorr value being approximately 50 μA·cm−2. This increase in porosity and weaker interfacial bonding, ultimately accelerated electrochemical corrosion. Therefore, a carefully balanced amount of HA significantly enhances the performance of the ZK60 magnesium alloy, while excessive amounts can be detrimental.