Earticle

본 논문에서는 인코더-디코더 모델(encoder-decoder model)에서 잠재 벡터(latent vector)의 분류 성능을 비교 분석한다. 오토인코더와 같은 일반적인 인코더-디코더 모델은 인코더 입력을 잠재 벡터로 변환하고 이를 디코더에 입력하여 인코더 입력과 유사한 출력을 생성하도록 학습한다. 이와 같은 인코더-디코더 모델의 잠재 벡터는 인코더 입력의 특징을 추상화하여 잘 보존한다고 고려할 수 있다. 나아가 잠재 벡터가 특징 공간에서 클러스터들 사이에서 구분이 가능한 거리를 보장한다면 이를 비지도 학습에 적용하는 것이 가능하다. 본 논문에서는 인코더-디코더 모델 에서의 잠재 벡터를 비지도 학습 및 점진적 학습에 적용하기 위한 기초 연구로서 잠재 벡터의 분류 성능을 분석한 다. 이를 위해 스택트 오토인코더(stacked autoencoder)와 2가지 종류의 CNN(Convolutional Neural Network) 기반 오토인코더를 바탕으로 각각 구해지는 잠재 벡터를 KNN(K-Nearest Neighbor)과 랜덤 포레스 트(random forest)를 포함하는 4가지 종류의 분류기에 적용한다. 실험 결과 완전 연결 계층(fully connected or dense layer)를 가지는 CNN 기반의 오토인코더를 사용한 결과 평균 정확률은 약 97%이고 스택트 오토인코더의 결과는 약 95%로 2% 정도 우수한 분류 성능을 보이는 것을 확인할 수 있다. 본 논문에서의 연구를 바탕으로 완전 연결 계층을 가지는 CNN 기반의 오토 인코더를 이용하여 구해지는 잠재 벡터를 비지도 학습에 적용하는 것으로 확 장하는 것이 가능하다.

This paper compares and analyzes the classification performance of latent vectors in the encoder-decoder model. A typical encoder-decoder model, such as an autoencoder, transforms the encoder input into a latent vector and feeds it into the decoder. In this process, the encoderdecoder model learns to produce an decoder output similar to the encoder input. We can consider that the latent vector of the encoder-decoder model is well preserved by abstracting the characteristics of the encoder input. Further, it is possible to apply to unsupervised learning, if the latent vector guarantees a sufficient distance between clusters in the feature space. In this paper, the classification performance of latent vectors is analyzed as a basic study for applying latent vectors in encoder-decoder models to unsupervised and continual learning. The latent vectors obtained by the stacked autoencoder and 2 types of CNN-based autoencoder are applied to 4 kinds of classifiers including KNN and random forest. Experimental results show that the latent vector using the CNN-based autoencoder with a dense layer(about 97%) shows superior classification performance by up to 2% compared to the result of the stacked autoencoder(about 95%). Based on the results in this paper, it is possible to extend the latent vector obtained by using a CNN-based auto-encoder with dense layer to unsupervised learning.

Three-dimensional (3D) point clouds provide detailed geometric understanding of real-world environments but remain challenging to process due to their sparse and unordered nature. Contrastive learning has emerged as a powerful self-supervised approach for learning representations from unlabeled 3D point cloud data. At the core of these methods lie encoder architectures that project raw points into discriminative latent spaces. This brief survey highlights major encoder families used in 3D contrastive learning and analyzes their design principles, strengths, and limitations. We further discuss how encoder choice influences downstream performance and outline research trends toward efficient, multimodal, and real-time contrastive frameworks.

A surface encoder based on the Half-shaded square pattern (HSSP) is presented. The HSSP working as reference grid is composed of the straight lines which are easy to be fabricated and make measuring time short. Since the periodic cell is separated in ON/OFF by the $45^{\circ}$ straight line, the duration from the starting point of scanning to the first rising edge and the duty cycle of the pulse train vary with respect to the position of the starting point. And the relationship between X and Y position and the duration, and duty cycle is described in the simple linear equation. Therefore, it is possible to measure X and Y position with the measured duration and duty cycle without calculating load. Through the test set-up, the feasibility of the proposed surface encoder was verified. Also the future works for improvement of performance were suggested.

Recently, neural architecture search (NAS) has received increasing attention as it can replace human experts in designing the architecture of neural networks for different tasks and has achieved remarkable results in many challenging tasks. In this study, a path-based computation neural architecture encoder (PCE) was proposed. Our PCE first encodes the computation of information on each path in a neural network, and then aggregates the encodings on all paths together through an attention mechanism, simulating the process of information computation along paths in a neural network and encoding the computation on the neural network instead of the structure of the graph, which is more consistent with the computational properties of neural networks. We performed an extensive comparison with eight encoding methods on two commonly used NAS search spaces (NAS-Bench-101 and NAS-Bench-201), which included a comparison of the predictive capabilities of performance predictors and search capabilities based on two search strategies (reinforcement learning-based and Bayesian optimization-based) when equipped with different encoders. Experimental evaluation shows that PCE is an efficient encoding method that effectively ranks and predicts neural architecture performance, thereby improving the search efficiency of neural architectures.

Imbalanced datasets are common in industrial internet of things (IIoT) systems due to challenges in acquiring faulty labels. Augmentation and graph-based methods have been proposed to improve classification accuracy of deep learning-based systems. However, the conventional approaches can be limited by training complexity and inefficient memory usage. In this paper, GraphSAGE with contrastive encoder (GCE) is proposed to improve classification accuracy and memory utilization efficiency. From the simulation results, it is confirmed that the GCE can improve classification accuracy by up to 23% compared to conventional approaches.

BCH code is one of the most widely used error correcting code for the detection and correction of random errors in the modern digital communication systems. The conventional BCH encoder that is operated in bit-serial manner cannot adequate with the recent high speed appliances. Therefore, parallel encoding algorithms are always a necessity. In this paper, we introduced a new systolic array type BCH parallel encoder. To study the area and speed, several parallel factors of the systolic array encoder is compared. Furthermore, to prove the efficiency of the proposed algorithm using tree-type structure, the throughput and the area overhead was compared with its counterparts also. The proposed BCH encoder has a great flexibility in parallelization and the speed was increased by 40% than the original one. The results were implemented on synthesis and simulation on FPGA using VHDL.

This paper describes a second-generation dual sine-grid surface encoder for 2-D position measurements. The surface encoder consisted of a 2-D grid with a 2-D sinusoidal pattern on its surface, and a 2-D angle sensor that detected the 2-D profile of the surface grid The 2-D angle sensor design of previously developed first-generation surface encoders was based on geometric optics. To improve the resolution of the surface encoder, we fabricated a 2-D sine-grid with a pitch of $10{\mu}m$. We also established a new optical model for the second-generation surface encoder that utilizes diffraction and interference to generate its measured values. The 2-D sine-grid was fabricated on a workpiece by an ultra precision lathe with the assistance of a fast tool servo. We then performed a UV-casting process to imprint the sine-grid on a transparent plastic film and constructed an experimental setup to realize the second-generation surface encoder. We conducted tests that demonstrated the feasibility of the proposed surface encoder model.

In this study, we explore the potential of leveraging machine learning techniques, specifically auto-encoders (AE), for the decoding of linear block codes. Our findings suggest that this approach can outperform the conventional ordered statistics decoding (OSD) method, especially in a Rayleigh fading channel environment. We have rigorously trained the AE under both additive white Gaussian noise and Rayleigh fading channel conditions to ensure robustness in its performance. The output of the AE is combined with the received vector in a suitable manner to perform OSD. Through our experiments, we demonstrate that this proposed decoding approach yields better results than the conventional OSD method in Rayleigh fading channel when we used (23,12) Golay code.

This paper proposes two new design approaches for decimal to binary-coded-decimal (BCD) encoder using reversible logic through Peres gate (PG) and Feynman gate (FG) which consume 10 and 11 gates respectively to realize such circuitry. Gates have been arranged properly to minimize the gate count (GC) followed by the reduction of the quantum cost (QC) and garbage outputs (GO). The proposed designs have an improvement at least up to 45%, 40%, and 50% of GC, GO and QC respectively from the best state of the art designs.

Music brings pleasure and relaxation to people. Therefore, it is necessary to classify musical genres based on scenes. Identifying favorite musical genres from massive music data is a time-consuming and laborious task. Recent studies have suggested that machine learning algorithms are effective in distinguishing between various musical genres. However, meeting the actual requirements in terms of accuracy or timeliness is challenging. In this study, a hybrid machine learning model that combines a deep residual auto-encoder (DRAE) and support vector machine (SVM) for musical genre recognition was proposed. Eight manually extracted features from the Mel-frequency cepstral coefficients (MFCC) were employed in the preprocessing stage as the hybrid music data source. During the training stage, DRAE was employed to extract feature maps, which were then used as input for the SVM classifier. The experimental results indicated that this method achieved a 91.54% F1-score and 91.58% top-1 accuracy, outperforming existing approaches. This novel approach leverages deep architecture and conventional machine learning algorithms and provides a new horizon for musical genre classification tasks.

One of the most common task of electromyography (EMG)-based human?machine interface (HMI) is the hand gesture recognition. Robust and accurate predictions leveraging the surface EMG (sEMG) is a key challenge. In this paper, we present a 3D hand gesture prediction application, leveraging the sEMG signal and the optical hand tracking information. A transformer?encoder classifier (TEC) module is introduced in an IPC-system to predict the 3D-hand gestures using eight monopolar channels from sEMG as input. An experimental testbed is setup to acquire, train, and predict the 3D hand gestures within a feasible range of performance. The performance has been evaluated in terms of percentage of correctly classified keypoints (PCK). PCK is measured by first estimating the euclidean distance between the actual and the predicted keypoints. The percentage of keypoints within a threshold distance value are then calculated. Results from the ablation study indicate that the proposed scheme shows a percentage of correctly classified keypoints of up to 72.8%, 92.7%, 97.2%, and 98.6% with a PCK threshold of 5 mm, 10 mm, 20 mm, and 30 mm, respectively.

Reliable channel estimation is critical for achieving high-speed and energy-efficient communication in Unmanned Aerial Vehicle-Free Space Optical (UAV-FSO) systems, particularly under dynamic impairments such as atmospheric turbulence (AT) and pointing errors (PEs). This paper proposes a pilot-free channel estimation framework based on a Variational Autoencoder (VAE). The system employs Intensity Modulation/Direct Detection (IM/DD) with?-ary one-hot encoded symbols (). The VAE encodes noisy received signals into a 128-dimensional latent space and reconstructs the transmitted data, effectively learning the joint effects of AT, PEs, and AWGN. Unlike prior works that primarily consider boresight or Gaussian-jitter PEs, this study explicitly incorporates a Nakagami-modeled PE distribution, capturing UAV-induced beam misalignment under mobility, vibration, and turbulence coupling. Simulation results show that the proposed VAE significantly outperforms conventional estimators (LS, MMSE, LMMSE) and deep learning baselines (AE, DNN, CNN) across various turbulence strengths. Under strong turbulence and PEs, the VAE attains nearly two-fold lower MSE compared to CNN and DNN. In addition, evaluation on real turbulence-impaired datasets further validates robustness and generalization. The proposed pilot-free scheme delivers accurate channel estimation, reduced BER, and improved spectral efficiency, making it suitable for real-time adaptive UAV-FSO communication.

일반적인 평판 프린터 시스템은 PC와 전용 S/W로 구성되어 사용에 불편함이 존재한다. 결국, 평판 프린터 시스템 구성의 간소화, 스마트화 등을 통해 다양한 형태의 프린팅을 쉽고 편리하게 사용할 수 있는 기술이 필요하 다. 즉, 한 대의 프린터로 여러 종류의 소재에 인쇄가 가능하며, 다양한 형태의 상품을 인쇄할 수 있는 다입체 프린 터에 대한 수요가 증가하고 있다. 따라서, 본 논문에서는 Head Encoder/Trigger 제어를 이용하여 다입체 인쇄가 가능한 평판 프린터 시스템을 개발하였다. 이를 위해, 평판 프린터 내부 모듈을 입력 형태 감지 센서를 연계시키고, 별도의 메인 콘트롤러를 통해 프린터의 Head Encoder와 Head Trigger 신호에 의해 모든 동작 상태를 제어하는 평판 프린터를 개발하였다. 이를 통해, IoT 기술의 발전 및 보급의 확산으로 산업 전반에 걸쳐 스마트 환경의 프린 터 제어가 발전된 형태로 확대될 것이며, 향후 3D 프린팅 산업 발전에 기여할 것으로 기대된다.

The general flatbed printer system is composed of a PC and a dedicated S/W, which is inconvenient to use. In the end, there is a need for a technology that can easily and conveniently use various types of printing through simplification, smartization, etc. of a flatbed printer system configuration. That is, there is an increasing demand for multi-dimensional printer capable of printing on various types of materials with one printer and capable of printing various types of products. Therefore, in this paper, we developed a flatbed printer system capable of multi-dimensional printing using Head Encoder/Trigger control. To this end, we developed a flatbed printer that connects the internal module of the flatbed printer with an input type detection sensor and controls all operating states by the head encoder and head trigger signals of the printer through separate main controllers. Through this, the development and diffusion of IoT technology will expand the printer control of the smart environment to the developed form throughout the industry. It is expected to contribute to the development of the 3D printing industry in the future.

최근 인공지능(AI) 기술의 발전으로 기존 침입 탐지 시스템(NIDS)의 한계를 극복하기 위한 연구가 활발히 이루어지고 있 다. 기존 서명 기반 또는 단일 딥러닝 모델은 새로운 공격 유형이나 제로데이 공격 탐지에 취약하며 네트워크 플로우 간 구조 적 관계를 충분히 반영하지 못한다. 이에 본 연구는 SAGEConv-GNN과 Transformer Encoder를 결합한 하이브리드 NIDS를 제안한다. GNN은 국부 구조 정보를, Transformer는 장기 의존성과 전역 문맥을 학습하며 Gated Fusion으로 이를 통합한다. 또 한 positional encoding, Early Stopping, label smoothing, class weight 등을 적용해 학습 안정성을 확보하였다. UNSW-NB15 (약 50,000 flows) 실험 결과 본 모델은 GCN, GAT, GraphSAGE, Transformer보다 우수했으며 Accuracy 99.70%, Precision 96.44%, Recall 99.08%, F1-score 97.2%를 달성했다. 이는 불균형 데이터 환경에서도 높은 탐지 성능과 낮은 오탐지율을 유지 함을 보이며 AI 기반 하이브리드 NIDS의 실용 가능성을 제시한다.

With recent advances in AI, overcoming the limits of traditional NIDS has become crucial: signature-based or single deep learning models struggle with novel/zero-day attacks and fail to capture structural relations among flows. We propose a hybrid NIDS combining SAGEConv-GNN (local structure) and a Transformer Encoder (global context/long-range dependencies), integrated via Gated Fusion, with positional encoding, early stopping, label smoothing, and class weighting for stable training. On UNSW-NB15 (~50k flows), our model outperforms GCN, GAT, GraphSAGE, and Transformer baselines, achieving 99.70% Accuracy, 96.44% Precision, 99.08% Recall, and 97.2% F1. This demonstrates strong detection performance and low false positives in imbalanced settings, underscoring the practicality of AI-based hybrid NIDS.

본 논문에서는 원격 측정 시스템 파라미터에 대한 실시간 업데이트가 가능한 PCM (pulse code modulation) 엔코더 구조에 대해 기술한다. PCM 엔코더 내부에는 FPGA (filed programmable gate array), flash 메모리, 센서 데이터 계측을 위한 아날로그 신호 조절부를 구성하였다. PCM 엔코더의 FPGA 내부에 로직을 통해 UART (universal asynchronous receiver/transmitter) 통신, 아날로그 신호 조절부 제어, flash 메모리 제어, 프레임 구성이 가능하다. UART 통신을 이용해 PC에서 파라미터 데이터를 PCM 엔코더에게 송신할 수 있으며, flash 메모리가 제어되어 원격 계측 시스템의 파라미터가 실시간으로 업데이트 되어 최종적으로 프레임이 구성된다. 시뮬레이션과 검증을 통해 파라미터 데이터의 실시간 업데이트 여부에 대해 확인하였으며, 제안된 구조를 이용하여 유연성과 편의성을 높인 원격 계측 시스템을 구성할 수 있음을 확인하였다.

In this paper, we describe a PCM encoder structure that can update the telemetry system parameters in real time. In the PCM encoder, an analog signal control unit for FPGA, flash memory, and sensor data acquisition was constructed. UART communication, analog signal control, flash memory control, and frame generation are possible through logic inside FPGA of PCM encoder. UART communication allows the PC to transmit parameter data to the PCM encoder, and flash memory is controlled to update the parameter of the telemetry system in real time and finally the frame is formed. Simulation and verification were performed to confirm whether the parameter data is updated in real time, and the proposed structure was used to construct a telemetry system with enhanced flexibility and convenience.

Incremental encoder is a common choice for measuring the rotational angle of an axis. It produces, however, some measurement errors, if some kinds of rotational vibration are occurred. Even by increasing the number of slits on the circular peripheral, these kinds of error can not be removed. It only decreases the probability of the error being generated by making the magnitude of encoder resolution smaller than that of the axial vibration. This paper investigates the error mechanism of incremental encoder and suggests a digital logic circuit that prevents the measurement error to happen.

The integrated chip of optical micro-encoder was fabricated and the feasibility as displacement measurement device was confirmed. The geometry of micro-encoder was designed to utilize the optical interference effect on the second order of diffracted beams. The hybrid-type micro-encoder consisted with light emitting diode, photodiode, polyimide wave-guide and micro-lens provides stable micro-encoding results for high speed displacements. The measurement shows the resolution of displacement of 1.00 +/- 0.02 ${\mu}m$ for the grating with scale pitch of 2.0${\mu}m$.

본 논문에서는 커뮤테이션 방식을 이용한 심플 프로그래머블 PCM 엔코더 구조를 제시하고 구현하였다. 텔레메트리 시스템은 센서에서 획득한 데이터들로 프레임을 생성하기 위해 각각의 데이터들을 채널에 할당할 수 있는 정보가 필요하다. 이때 상태 정보의 수가 많거나 데이터 타입이 다양할 경우, 각 채널에 대량의 정보를 입력해야 할 필요성이 존재한다. 그러나 채널수와 데이터양이 많을수록 오류가 발생할 가능성이 증가한다. 따라서 본 논문에서는 프로그램을 이용하여 채널 정보를 작성하고, ROM에 채널 정보를 저장할 수 있는 PCM 엔코더를 구현하였다. 본 논문에서 제안한 PCM 엔코더 구조는 오류의 발생 가능성을 줄이고, 채널 정보 입력 소스 코드 길이의 축소 등 개발 속도를 향상시킬 수 있다. 또한 시뮬레이션을 이용하여 제안한 구조의 타당성을 확인하였다.

This paper presents and implements a simple programmable PCM encoder structure uisng the commutation method. In the telemetry system, information is required to assign each data to the channel in order to generate a frame format the data acpuired from the sensor. In this case, when the number of state information is large or the data type is various, there is a necessity to input a large amount of information to each channel. However, the more the number of channels and data, the more probability the error will occur. Therefore, in this paper, the channel information is created using the program. And PCM encoder was implemented to store channel information in ROM. The proposed PCM encoder architecture reduces the likelihood of errors. And it can improve the development speed. The validity of proposed structure is proved by simulation.