News

Most Read

Published in last 1 year |  In last 2 years |  In last 3 years |  All
Please wait a minute...
For Selected: Toggle Thumbnails
Graph convolutional network combined with random walks and graph attention network for node classification
Chen Yong, Xie Xiaozhu, Weng Wei
The Journal of China Universities of Posts and Telecommunications    2024, 31 (3): 1-14.   DOI: 10.19682/j.cnki.1005-8885.2024.1004
Abstract307)      PDF(pc) (1823KB)(41)       Save
Graph conjoint attention (CAT) network is one of the best graph convolutional networks (GCNs) frameworks,
which uses a weighting mechanism to identify important neighbor nodes. However, this weighting mechanism is
learned based on static information, which means it is susceptible to noisy nodes and edges, resulting in significant
limitations. In this paper, a method is proposed to obtain context dynamically based on random walk, which allows
the context-based weighting mechanism to better avoid noise interference. Furthermore, the proposed context-based
weighting mechanism is combined with the node content-based weighting mechanism of the graph attention (GAT)
network to form a model based on a mixed weighting mechanism. The model is named as the context-based and
content-based graph convolutional network (CCGCN). CCGCN can better discover important neighbors, eliminate
noise edges, and learn node embedding by message passing. Experiments show that CCGCN achieves state-of-the-
art performance on node classification tasks in multiple datasets.
Reference | Related Articles | Metrics
Black-box membership inference attacks based on shadow model
Han Zhen, Zhou Wen'an, Han Xiaoxuan, Wu Jie
The Journal of China Universities of Posts and Telecommunications    2024, 31 (4): 1-16.   DOI: 10.19682/j.cnki.1005-8885.2024.1016
Abstract290)      PDF(pc) (3603KB)(140)       Save
Membership inference attacks on machine learning models have drawn significant attention. While current  research primarily utilizes shadow modeling techniques, which require knowledge of the target model and training  data, practical scenarios involve black-box access to the target model with no available information. Limited  training data further complicate the implementation of these attacks. In this paper, we experimentally compare  common data enhancement schemes and propose a data synthesis framework based on the variational autoencoder  generative adversarial network (VAE-GAN) to extend the training data for shadow models. Meanwhile, this paper  proposes a shadow model training algorithm based on adversarial training to improve the shadow model's ability to  mimic the predicted behavior of the target model when the target model's information is unknown. By conducting  attack experiments on different models under the black-box access setting, this paper verifies the effectiveness of the  VAE-GAN-based data synthesis framework for improving the accuracy of membership inference attack.  Furthermore, we verify that the shadow model, trained by using the adversarial training approach, effectively  improves the degree of mimicking the predicted behavior of the target model. Compared with existing research  methods, the method proposed in this paper achieves a 2% improvement in attack accuracy and delivers better  attack performance.
Reference | Related Articles | Metrics
Personalized trajectory data perturbation algorithm based on quadtree indexing
The Journal of China Universities of Posts and Telecommunications    2024, 31 (4): 17-27.   DOI: 10.19682/j.cnki.1005-8885.2024.1014
Abstract286)      PDF(pc) (1357KB)(46)       Save
To solve the privacy leakage problem of truck trajectories in intelligent logistics, this paper proposes a Quadtree-based Personalized Joint location Perturbation (QPJLP) algorithm using location generalization and local differential privacy techniques. Firstly, a flexible position encoding mechanism based on the spatial quadtree indexing is designed, and the length of the encoding can be adjusted freely according to data availability. Secondly, to meet the privacy needs of different locations of users, location categories are introduced to classify locations as sensitive and ordinary locations. Finally, the truck invokes the corresponding mechanism in the QPJLP algorithm to locally perturb the code according to the location category, allowing the protection of non-sensitive locations to be reduced without weakening the protection of sensitive locations, thereby improving data availability. Simulation experiments demonstrate that the proposed algorithm effectively meets the personalized trajectory privacy requirements while also exhibiting good performance in trajectory proportion estimation and Top-K classification.
Reference | Related Articles | Metrics
Superjunction 4H-SiC trench-gate IGBT with an integrated clamping PN diode
The Journal of China Universities of Posts and Telecommunications    2024, 31 (2): 3-9.   DOI: 10.19682/j.cnki.1005-8885.2024.0001
Abstract280)      PDF(pc) (4079KB)(136)       Save

In this paper, a novel superjunction 4H-silicon carbide (4H-SiC) trench-gate insulated-gate bipolar transistor (IGBT) featuring an integrated clamping PN diode between the P-shield and emitter (TSD-IGBT) is designed and theoretically studied. The heavily doping superjunction layer contributes to a low specific on-resistance, excellent electric field distribution, and quasi-unipolar drift current. The anode of the clamping diode is in floating contact with the P-shield. In the on-state, the potential of the P-shield is raised to the turn-on voltage of the clamping diode, which prevents the hole extraction below the N-type carrier storage layer (NCSL). Additionally, during the turn-off transient, once the clamping diode is turned on, it also promotes an additional hole extraction path. Furthermore, the potential dropped at the semiconductor near the trench-gate oxide is effectively reduced in the off-state.

Reference | Related Articles | Metrics
Performance study of vertical MSM solar-blind photodetectors based on β-Ga 2O 3 thin film
The Journal of China Universities of Posts and Telecommunications    2024, 31 (2): 17-27.   DOI: 10.19682/j.cnki.1005-8885.2024.0006
Abstract280)      PDF(pc) (4363KB)(123)       Save

In this work, β-Ga2O3 thin films were grown on SiO2 substrate by atomic layer deposition (ALD) and annealed in N2 atmosphere to enhance the crystallization quality of the thin films, which were verified from X-rays diffraction (XRD). Based on the grown β-Ga2O3 thin films, vertical metal-semiconductor-metal (MSM) interdigital photodetectors (PDs) were fabricated and investigated. The PDs have an ultralow dark current of 1.92 pA, ultra-high photo-to-dark current ratio (PDCR) of 1.7×106, and ultra-high detectivity of 4.25×1014 Jones at a bias voltage of 10 V under 254 nm deep ultraviolet (DUV). Compared with the horizontal MSM PDs under the same process, the PDCR and detectivity of the fabricated vertical PDs are increased by 1 000 times and 100 times, respectively. In addition, the vertical PDs possess a high responsivity of 34.24 A/W and an external quantμm efficiency of 1.67×104%, and also exhibit robustness and repeatability, which indicate excellent performance. Then the effects of electrode size and external irradiation conditions on the performance of the vertical PDs continued to be investigated.

Reference | Related Articles | Metrics
Artificial rabbit optimization algorithm based on chaotic mapping and Levy flight improvement
The Journal of China Universities of Posts and Telecommunications    2024, 31 (4): 54-69.   DOI: 10.19682/j.cnki.1005-8885.2024.1010
Abstract270)      PDF(pc) (2941KB)(31)       Save
An artificial rabbit optimization algorithm based on chaotic mapping and Levy flight improvement is proposed, which has the advantages of good initial population quality and fast convergence compared with the traditional artificial rabbit optimization algorithm, called CLARO. CLARO’s improvement method starts from three aspects: to optimize the quality of the initial population of the algorithm a chaotic mapping is brought in to initialize the population; to avoid the algorithm from falling into local optimum Levy flight is added in the exploration phase and the threshold of energy factor A is optimized to better balance exploration and exploitation. The efficiency of CLARO is tested on a set of 23 benchmark function sets by comparing it with ARO and different meta-heuristics algorithms. At last, the comparison experiments conclude that all three improvement strategies enhance the performance of ARO to some extent, with Levy flight providing the most significant improvement in ARO performance. The experimental results showed that CLARO has better results and faster convergence compared to other algorithms, while successfully addressing the drawbacks of ARO and being able to face more challenging problems.
Reference | Related Articles | Metrics

Trench gate GaN IGBT with controlled hole injection efficiency

The Journal of China Universities of Posts and Telecommunications    2024, 31 (2): 10-16.   DOI: 10.19682/j.cnki.1005-8885.2024.0012
Abstract265)      PDF(pc) (2095KB)(102)       Save

In this paper, a novel trench gate gallium nitride insulated gate bipolar transistor (GaN IGBT), in which the collector is divided into multiple regions to control the hole injection efficiency, is designed and theoretically studied. The incorporation of a P+/P- multi-region alternating structure in the collector region mitigates hole injection within the collector region. When the device is in forward conduction, the conductivity modulation effect results in a reduced storage of carriers in the drift region. As a result, the number of carriers requiring extraction during device turn-off is minimized, leading to faster turn-off speed. The results illustrate that the GaN IGBT with controlled hole injection efficiency (CEH GaN IGBT) exhibits markedly enhanced performance compared to conventional GaN IGBT, showing a remarkable 42.2% reduction in turn-off time and a notable 28.5% decrease in turn-off loss.

Reference | Related Articles | Metrics
Power and Rate Adaption in Wireless Communication Systems with Energy Harvesting–Based on Soft Decision Decoding
The Journal of China Universities of Posts and Telecommunications    2024, 31 (4): 70-82.   DOI: 10.19682/j.cnki.1005-8885.2024.1017
Abstract247)      PDF(pc) (2041KB)(37)       Save
In this paper, the online power control and rate adaptation for a wireless communication system with energy harvesting are investigated, in which soft decision decoding is adopted by the receiver. To efficiently utilize the harvested energy and maximize the average actual information transmission rate, transmit power, modulation order and code rate are jointly optimized. The Lyapunov framework is utilized to transform the long-term optimization problem into an optimization problem per time slot. Since there is no theoretical formula for the error rate of soft decision decoding, the optimization problem cannot be analytically solved. A table to find the optimal modulation order and code rate under the different values of signal-to-noise ratio is built first, and then a numeric algorithm to find the solution to the optimization problem is given. The feasibility and performance of the proposed algorithm are demonstrated by simulation. The simulation results show that compared with the algorithms to maximize the theoretic channel capacity, the proposed algorithm can achieve a higher actual transmission rate.
Reference | Related Articles | Metrics
Improving Link Prediction Models through a Performance Enhancement Scheme: A Study on Semi-Supervised Learning and Model Soup
The Journal of China Universities of Posts and Telecommunications    2024, 31 (4): 43-53.   DOI: 10.19682/j.cnki.1005-8885.2024.1015
Abstract245)      PDF(pc) (2574KB)(40)       Save
As a fact, most constructed knowledge graphs are far from complete no matter its size. This incompleteness will cause negative influence on the applications based on knowledge graphs. As an important method for knowledge graph complementation, link prediction has become a hot research topic in recent years. In this paper, a performance enhancement scheme for link prediction models based on the idea of semi-supervised learning and model soup is proposed, which effectively improves the model performance on several mainstream link prediction models with small changes to their architecture. This novel scheme consists of two main parts: (1) predicting potential fact triples in the graph with semi-supervised learning strategies, (2) creativily combining semi-supervised learning and model soup to further improve the final model performance without adding significant computational overhead. We experimentally validate the effectiveness of the scheme for a variety of link prediction models, especially on the dataset with dense relationships. In terms of CompGCN, the model with the best overall performance among the tested models improves its Hits@1 metric by 14.7% on the FB15K-237 dataset and 7.8% on the WN18RR dataset after using the enhancement scheme. Meanwhile, we observe that the semi-supervised learning strategy in the augmentation scheme has significant improvement for multi-class link prediction models, and the performance improvement brought by the introduction of the model soup is related to the specific tested models, because performance of some models are improved while others remained largely unaffected.
Reference | Related Articles | Metrics
Dynamic coverage of mobile multi-target in sensor networks based on virtual force
The Journal of China Universities of Posts and Telecommunications    2024, 31 (4): 83-94.   DOI: 10.19682/j.cnki.1005-8885.2024.1006
Abstract223)      PDF(pc) (3233KB)(26)       Save
A new procedure of distributed self-control coverage for monitoring the dynamic targets with mobile sensor network is proposed. A special model is given for maintaining the nodes bi-connectivity and optimizing the coverage of the moving targets. The model consists of two parts, the virtual force model is proposed for motion control and the whale optimization algorithm is improved to further optimize the node positions and to reach the steady state quickly. The virtual resultant force stretches the network toward the uncovered targets by its multi-target attractive force, and maintains the network connectivity by its attractive force while network stretching, and avoids node collisions by its repulsive force while nodes moving. The operating mechanism of multi-target attractive force and other forces is also profoundly anatomized. The adjustment criteria for the model in different application scenarios are also given. Finally, the comparisons are shown to be significant advantages over other similar kinds.
Reference | Related Articles | Metrics
Deep kernel extreme learning machine classifier based on the improved sparrow search algorithm
Zhao Guangyuan, Lei Yu
The Journal of China Universities of Posts and Telecommunications    2024, 31 (3): 15-29.   DOI: 10.19682/j.cnki.1005-8885.2024.1003
Abstract222)      PDF(pc) (3906KB)(34)       Save
In the classification problem, deep kernel extreme learning machine (DKELM) has the characteristics of efficient
processing and superior performance, but its parameters optimization is difficult. To improve the classification
accuracy of DKELM, a DKELM algorithm optimized by the improved sparrow search algorithm (ISSA), named as
ISSA-DKELM, is proposed in this paper. Aiming at the parameter selection problem of DKELM, the DKELM
classifier is constructed by using the optimal parameters obtained by ISSA optimization. In order to make up for the
shortcomings of the basic sparrow search algorithm (SSA), the chaotic transformation is first applied to initialize the
sparrow position. Then, the position of the discoverer sparrow population is dynamically adjusted. A learning
operator in the teaching-learning-based algorithm is fused to improve the position update operation of the joiners.
Finally, the Gaussian mutation strategy is added in the later iteration of the algorithm to make the sparrow jump out
of local optimum. The experimental results show that the proposed DKELM classifier is feasible and effective, and
compared with other classification algorithms,the proposed DKELM algorithm aciheves better test accuracy.
Reference | Related Articles | Metrics
All-dielectric terahertz metasurface governed by bound states in the continuum with high-Q factor
The Journal of China Universities of Posts and Telecommunications    2024, 31 (2): 44-54.   DOI: 10.19682/j.cnki.1005-8885.2024.0003
Abstract221)      PDF(pc) (3460KB)(55)       Save
The method of terahertz (THz) resonance with a high-quality (high-Q) factor offers a vital physical mechanism for metasurface sensors and other high-Q factor applications. However, it is challenging to excite the resonance with a high-Q factor in metasurfaces with proper sensitivity as well as figure of merit (FOM) values. Here,  an all-dielectric metasurface composed of two asymmetrical rectangular blocks is suggested. Quartz and silicon are the materials applied for the substrate and cuboids respectively. The distinct resonance governed by bound states in the continuum (BIC) is excited by forming an asymmetric cluster by a novel hybrid method of cutting and moving the cuboids. The investigation focuses on analyzing the transmission spectra of the metasurface under different variations in structural parameters and the loss of silicon refractive index. When the proposed defective metasurface serves as a transmittance sensor, it shows a Q factor of 1.08×10 4 and achieves a FOM up to 4.8×10 6, which is obtained under the asymmetric parameter equalling 1 μm. Simultaneously, the proposed defective metasurface is sensitive to small changes in refractive index. When the thickness of the analyte is 180 μm, the sensitivity reaches a maximum value of 578 GHz / RIU. Hence, the proposed defective metasurface exhibits an extensive number of possible applications in the filters, biomedical diagnosis, security screening, and so on.
Reference | Related Articles | Metrics
Design of digital calibration based on variable step size of sub-binary SAR ADC
The Journal of China Universities of Posts and Telecommunications    2024, 31 (2): 62-71.   DOI: 10.19682/j.cnki.1005-8885.2024.00010
Abstract220)      PDF(pc) (5060KB)(63)       Save
Addressing the impact of capacitor mismatch on the conversion accuracy of successive approximation register analog-to-digital converter (SAR ADC), a 12-bit 1 MS/s sub-binary SAR ADC designed using variable step size digital calibration was proposed. The least mean square (LMS) calibration algorithm was employed with a ramp signal used as the calibration input signal. Weight errors, extracted under injected disturbances, underwent iterative training to optimize weight values. To address the trade-off between conversion accuracy and speed caused by a fixed step size,  a novel variable step size algorithm tailored for SAR ADC calibration was propased. The core circuit and layout of the SAR ADC were implemented using the Taiwan Semiconductor manufacturing Company (TSMC) 0.35 μm complementary metal-oxide-semiconductor (CMOS) commercial process. Simulation of the SAR ADC calibration algorithm was conducted using Simulink, demonstrating quick convergence and meeting conversion accuracy requirements compared to fixed step size simulation. The results indicated that the convergence speed of the LMS digital calibration algorithm with variable step size was approximately eight times faster than that with a fixed step size, also yielding a lower mean square error (MSE). After calibration, the simulation results for the SAR ADC exhibited an effective number of bit (ENOB) of 11.79 bit and a signal-to-noise and distortion ratio (SNDR) of 72.72 dB, signifying a notable enhancement in the SAR ADC performance.
Reference | Related Articles | Metrics
Design and implementation of a multi-tile parallel scanning rasterization accelerator
The Journal of China Universities of Posts and Telecommunications    2024, 31 (2): 94-104.   DOI: 10.19682/j.cnki.1005-8885.2024.0009
Abstract212)      PDF(pc) (6311KB)(58)       Save
In the design of a graphic processing unit (GPU), the processing speed of triangle rasterization is an important factor that determines the performance of the GPU. An architecture of a multi-tile parallel-scan rasterization accelerator was proposed in this paper. The accelerator uses a bounding box algorithm to improve scanning efficiency. It rasterizes multiple tiles in parallel and scans multiple lines at the same time within each tile. This highly parallel approach drastically improves the performance of rasterization. Using 65nm process standard cell library of Semiconductor Manufacturing International Corporation (SMIC), the accelerator can be synthesized to a maximum clock frequency of 220MHz. An implementation on the Genesys2 field programmable gate array (FPGA) board fully verifies the functionality of the accelerator. The implementation shows a significant improvement in rendering speed and efficiency and proves its suitability for high- performance rasterization.
Reference | Related Articles | Metrics
  Preparation and characteristic study of Schottky diodes based on Ga2O3 thin films
The Journal of China Universities of Posts and Telecommunications    2024, 31 (2): 28-37.   DOI: 10.19682/j.cnki.1005-8885.2024.0007
Abstract205)      PDF(pc) (3625KB)(79)       Save
This study uses atomic layer deposition (ALD) to grow Ga 2O 3 films on SiO 2 substrates and investigates the influence of film thickness and annealing temperature on film quality. Schottky diode devices are fabricated based on the grown Ga 2O 3 films, and the effects of annealing temperature, electrode size, and electrode spacing on the electrical characteristics of the devices are studied. The results show that as the film thickness increases, the breakdown voltage of the fabricated devices also increases. A Schottky diode with a thickness of 240 nm can achieve a reverse breakdown voltage of 300 V. The film quality significantly improves as the annealing temperature of the film increases. At a voltage of 5 V, the current of the film annealed at 900°C is 64 times that of the film annealed at 700°C. The optimum annealing temperature for Ohmic contact electrodes is 450°C. At 550°C, the Ohmic contact metal tends to burn, and the performance of the device is reduced. Reducing the electrode spacing increases the forward current of the device but decreases the reverse breakdown voltage. Increasing the Schottky contact electrode size increases the forward current, but the change is not significant, and there is no significant change in the reverse breakdown voltage. The device also performs well at high temperatures, with a reverse breakdown voltage of 220 V at 125°C.
Reference | Related Articles | Metrics
Energy-efficient reconfigurable processor for QC-LDPC via adaptive coding-voltage-frequency tuning
The Journal of China Universities of Posts and Telecommunications    2024, 31 (2): 72-84.   DOI: 10.19682/j.cnki.1005-8885.2024.0005
Abstract198)      PDF(pc) (3255KB)(65)       Save
To apply a quasi-cyclic low density parity check (QC-LDPC) to different scenarios, a data-driven pipelined macro-instruction set and a reconfigurable processor architecture are proposed for the typical QC-LDPC algorithm. The data-level parallelism is improved by instructions to dynamically configuring the multi-core computing units. Simultaneously, an intelligent adjustment strategy based on programmable wake-up controller (WuC) is designed so that the computing mode, operating voltage, and frequency of the QC-LDPC algorithm can be adjusted. This adjustment can improve the computing efficiency of the processor. The QC-LDPC decoders are verified on the Xilinx ZCU102 Field Programmable Gate Array (FPGA) board and the computing efficiency is measured. The experimental results indicate that the QC-LDPC processor can support two encoding lengths of three typical QC-LDPC algorithms and 20 adaptive operating modes of operating voltage and frequency. The maximum efficiency can reach up to 12.18 Mbit(s·mW) -1, which is more flexible than existing state-of-the-art processor for QC-LDPC.
Reference | Related Articles | Metrics
50-110 GHz, high isolation, and high-power linearity single-pole double-throw switch utilizing  100-nm GaN HEMT technology
The Journal of China Universities of Posts and Telecommunications    2024, 31 (2): 38-43.   DOI: 10.19682/j.cnki.1005-8885.2024.0011
Abstract190)      PDF(pc) (2721KB)(65)       Save

This article presents the design and performance of a single-pole double-throw (SPDT) switch operating in 50–110 GHz. The switch is fabricated in a 100-nm GaN high-electron-mobility transistors(HEMT) technology. To realize high-power capability, the dimensions of GaN HEMTs are selected by simulation verification. To enhance the isolation, an improved structure of shunt HEMT with two ground holes is employed. To extend the operation bandwidth, the SPDT switch with multi section resonant units is proposed and analyzed. To verify the SPDT switch design, a prototype operating in 50–110 GHz is fabricated. The measured results show that the fabricated SPDT switch monolithic microwave integrated circuit (MMIC) achieves an input 1 dB compression point (P1dB) of 38 dBm at 94 GHz, and an isolation within the range of 33 dB to 54 dB in 50–110 GHz. The insertion loss of the switch is less than 2.1 dB, while the voltage standing wave ratios (VSWR) of the input port and output port are both less than 1.8 in the operation bandwidth. Based on the measured results, the presented SPDT switch MMIC demonstrates high power capability and high isolation compared with other reported millimeter-wave SPDT MMIC designs.

Reference | Related Articles | Metrics
Convolutional neural network adaptation and optimization method in SIMT computing mode
zhenfu Feng Ya-Ying ZHANG Lele Yang Li-Dong XING
The Journal of China Universities of Posts and Telecommunications    2024, 31 (2): 105-112.   DOI: 10.19682/j.cnki.1005-8885.2024.0008
Abstract187)      PDF(pc) (3090KB)(59)       Save
For studying and optimizing the performance of general-purpose computing on graphics processing units(GPGPU) based on single instruction multiple threads(SIMT) processor about the neural network application, this work contributes a self-developed SIMT processor named Pomelo and correlated assembly program. The parallel mechanism of SIMT computing mode and self-developed Pomelo processor is briefly introduced. A common convolutional neural network(CNN) is built to verify the compatibility and functionality of the Pomelo processor. CNN computing flow with task level and hardware level optimization is adopted on the Pomelo processor. A specific algorithm for organizing a Z-shaped memory structure is developed, which addresses reducing memory access in mass data computing tasks. Performing the above-combined adaptation and optimization strategy, the experimental result illustrates that reducing memory access in SIMT computing mode plays a crucial role in improving performance. A 6.52 times performance is achieved on 4 processing elements case.
Reference | Related Articles | Metrics
CNN demodulation model with cascade parallel crossing  for CPM signals
Yang Jiachen, Duan Ruifeng, Li Chengju
The Journal of China Universities of Posts and Telecommunications    2024, 31 (3): 30-42.   DOI: 10.19682/j.cnki.1005-8885.2024.1005
Abstract187)      PDF(pc) (2056KB)(43)       Save
The continuous phase modulation (CPM) technique is widely used in range telemetry due to its high spectral
efficiency and power efficiency. However, the demodulation performance of the traditional maximum likelihood
sequence detection (MLSD) algorithm significantly deteriorates in non-ideal synchronization or fading channels. To
address this issue, this work proposes a convolutional neural network (CNN) called the cascade parallel crossing
network (CPCNet) to enhance the robustness of CPM signals demodulation. The CPCNet model employs a multiple
parallel structure and feature fusion to extract richer features from CPM signals. This approach constructs feature
maps at different levels, resulting in a more comprehensive training of the model and improved demodulation
performance. Simulation results show that under Gaussian channel, the proposed CPCNet achieves the same bit
error rate (BER) performance as MLSD method when there is no timing error, but with 1/4 symbol period timing
error, the proposed method has 2 dB demodulation gain compared with CNN and convolutional long short-term
memory deep neural network (CLDNN). In addition, under Rayleigh channel, the BER of the proposed method is
reduced by 5% -87% compared to that of MLSD in the wide signal-to-noise ratio (SNR) region.
Reference | Related Articles | Metrics
Predicting stability of integrated circuit test equipment using  upper side boundary values of normal distribution
The Journal of China Universities of Posts and Telecommunications    2024, 31 (2): 85-93.   DOI: 10.19682/j.cnki.1005-8885.2024.0002
Abstract184)      PDF(pc) (473KB)(61)       Save
In response to the growing complexity and performance of Integrated Circuit (IC), there is an urgent need to enhance the testing and stability of IC test equipment. A method was proposed to predict equipment stability using the upper side boundary value of normal distribution. Initially, the K-means clustering algorithm classifies and analyzes sample data. The accuracy of this boundary value is compared under two common confidence levels to select the optimal threshold. A range is then defined to categorize unqualified test data. Through experimental verification, the method achieves the purpose of measuring the stability of qualitative IC equipment through a deterministic threshold value and judging the stability of the equipment by comparing the number of unqualified data with the threshold value, which realizes the goal of long-term operation monitoring and stability analysis of IC test equipment.
Reference | Related Articles | Metrics