• The Journal of the Korea institute of electronic communication sciences
• /
• v.19 no.2
• /
• pp.417-426
• /
• 2024

In general, the implementation of machine learning requires prior knowledge and experience with deep learning models, and substantial computational resources and time are necessary for data processing. As a result, machine learning encounters several limitations when deployed on embedded processors. To address these challenges, this paper introduces a novel approach where a genetic algorithm is applied to the convolution operation within the machine learning process, specifically for performing a selective convolution operation.In the selective convolution operation, the convolution is executed exclusively on pixels identified by a genetic algorithm. This method selects and computes pixels based on a ratio determined by the genetic algorithm, effectively reducing the computational workload by the specified ratio. The paper thoroughly explores the integration of genetic algorithms into machine learning computations, monitoring the fitness of each generation to ascertain if it reaches the target value. This approach is then compared with the computational requirements of existing methods.The learning process involves iteratively training generations to ensure that the fitness adequately converges.

• IEMEK Journal of Embedded Systems and Applications
• /
• v.19 no.3
• /
• pp.151-158
• /
• 2024

Although on-device artificial intelligence (AI) has gained attention to diagnosing machine faults in real time, most previous studies did not consider the model retraining and redeployment processes that must be performed in real-world industrial environments. Our study addresses this challenge by proposing an on-device AI-based real-time machine fault diagnosis system that utilizes continual learning. Our proposed system includes a lightweight convolutional neural network (CNN) model, a continual learning algorithm, and a real-time monitoring service. First, we developed a lightweight 1D CNN model to reduce the cost of model deployment and enable real-time inference on the target edge device with limited computing resources. We then compared the performance of five continual learning algorithms with three public bearing fault datasets and selected the most effective algorithm for our system. Finally, we implemented a real-time monitoring service using an open-source data visualization framework. In the performance comparison results between continual learning algorithms, we found that the replay-based algorithms outperformed the regularization-based algorithms, and the experience replay (ER) algorithm had the best diagnostic accuracy. We further tuned the number and length of data samples used for a memory buffer of the ER algorithm to maximize its performance. We confirmed that the performance of the ER algorithm becomes higher when a longer data length is used. Consequently, the proposed system showed an accuracy of 98.7%, while only 16.5% of the previous data was stored in memory buffer. Our lightweight CNN model was also able to diagnose a fault type of one data sample within 3.76 ms on the Raspberry Pi 4B device.

• The Journal of the Acoustical Society of Korea
• /
• v.25 no.7
• /
• pp.339-345
• /
• 2006

We suggest new 2D two-way Parabolic equation algorithm for multiple scattering. Our method is based on the successive performance of the single scattering approach. First. as the single scattering algorithm, the reflected and transmitted fields are calculated at the vertical interface of a range independent sector. Then. the reflected field is saved and the transmitted field Propagated to the next vertical interface with the split-step Pade method. After one step ends, the same Process is repeatedly performed with the change of the Propagation direction until the reflected field at the vertical interface is close to zero. Final incoming and outgoing fields are obtained as the sum of the wave fields obtained for each step. Our algorithm is relatively simple for the numerical implementation and requires less computational resources than the existing algorithm for multiple scattering

• The Journal of the Acoustical Society of Korea
• /
• v.28 no.8
• /
• pp.815-821
• /
• 2009

Information processing in variable and noisy environments is usually accomplished by means of adaptive filters. Among various adaptive algorithms, Least Mean Square (LMS) has become the most popular for its robustness, good tracking capabilities and simplicity, both in terms of computational load and easiness of implementation. In practical application of the LMS algorithm, the most important key parameter is the Step Size. As is well known, if the Step Size is large, the convergence rate of the algorithm will be rapid, but the steady state mean square error (MSE) will increase. On the other hand, if the Step Size is small, the steady state MSE will be small, but the convergence rate will be slow. Many researches have been proposed to alleviate this drawback by using a variable Step Size. In this paper, a new variable Step Size LMS(VSSLMS) called Categorized VSSLMS (CVSSLMS) is proposed. CVSSLMS updates the Step Size by categorizing the current status of the gradient, hence significantly improves the convergence rate. The performance of the proposed algorithm was verified from the view point of convergence rate, Excessive Mean Square Error(EMSE), and complexity through experiments.

• Geomechanics and Engineering
• /
• v.36 no.2
• /
• pp.111-119
• /
• 2024

In the realm of rock excavation projects, precise estimation of the drilling rate index stands as a pivotal factor in strategic planning and cost assessment. This study introduces and evaluates two pioneering computational intelligence models designed for the prognostication of the drilling rate index, a pivotal parameter with direct implications for cost estimation in rock excavation projects. These models, denoted as the Relevance Vector Regression (RVR) optimized with the Invasive Weed Optimization algorithm (IWO) (RVR-IWO model) and the RVR integrated with the Shuffled Frog Leaping algorithm (SFL) (RVR-SFL model), represent a groundbreaking approach to forecasting drilling rate index. The RVR-IWO and RVR-SFL models were meticulously devised to harness the capabilities of computational intelligence and optimization techniques for drilling rate index estimation. This research pioneers the integration of IWO and SFL with RVR, constituting an unprecedented effort in forecasting drilling rate index. The primary objective of this study was to gauge the precision and dependability of these models in forecasting the drilling rate index, revealing significant distinctions between the two. In terms of predictive precision, the RVR-IWO model emerged as the superior choice when compared to the RVR-SFL model, underscoring the remarkable efficacy of the Invasive Weed Optimization algorithm. The RVR-IWO model delivered noteworthy results, boasting a Variance Account for (VAF) of 0.8406, a Mean Squared Error (MSE) of 0.0114, and a Squared Correlation Coefficient (R²) of 0.9315. On the contrary, the RVR-SFL model exhibited slightly lower precision, yielding an MSE of 0.0160, a VAF of 0.8205, and an R2 of 0.9120. These findings serve to highlight the potential of the RVR-IWO model as a formidable instrument for drilling rate index prediction, particularly within the framework of rock excavation projects. This research not only makes a significant contribution to the realm of drilling engineering but also underscores the broader adaptability of the RVR-IWO model in tackling an array of challenges within the domain of rock engineering. Ultimately, this study advances the comprehension of drilling rate index estimation and imparts valuable insights into the practical implementation of computational intelligence methodologies within the realm of engineering projects.

• Journal of the Institute of Electronics Engineers of Korea SD
• /
• v.40 no.3
• /
• pp.43-53
• /
• 2003

Unlike in analog display devices, the physical screen resolution in digital devices are fixed from the manufacturing. It is a weak point on digital devices. The screen resolution displayed in digital display devices is varied. Thus, interpolation or decimation of the resolution on the display is needed to make the input pixels equal to the screen resolution., This process is called image scaling. Many researches have been developed to reduce the hardware cost and distortion of the image of image scaling algorithm. In this paper, we proposed a Winscale algorithm. which modifies the scale up/down in continuous domain to the scale up/down in discrete domain. Thus, the algorithm is suitable to digital display devices. Hardware implementation of the image scaler is performed using Verilog XL and chip is fabricated in a 0.5${\mu}{\textrm}{m}$ Samsung SOG technology. The hardware costs as well as the scalabilities are compared with the conventional image scaling algorithms that are used in other software. This Winscale algorithm is proved more scalable than other image-scaling algorithm, which has similar H/W cost. This image-scaling algorithm can be used in various digital display devices that need image scaling process.

• Journal of the Institute of Electronics Engineers of Korea SD
• /
• v.45 no.6
• /
• pp.49-59
• /
• 2008

SIFT(Scale Invariant Feature Transform) is an algorithm to extract vectors at pixels around keypoints, in which the pixel colors are very different from neighbors, such as vortices and edges of an object. The SIFT algorithm is being actively researched for various image processing applications including 3-D image constructions, and its most computation-intensive stage is a keypoint localization. In this paper, we develope a fixed-point model of the keypoint localization and propose its efficient hardware architecture for embedded applications. The bit-length of key variables are determined based on two performance measures: localization accuracy and error rate. Comparing with the original algorithm (implemented in Matlab), the accuracy and error rate of the proposed fixed point model are 93.57% and 2.72% respectively. In addition, we found that most of missing keypoints appeared at the edges of an object which are not very important in the case of keypoints matching. We estimate that the hardware implementation will give processing speed of $10{\sim}15\;frame/sec$, while its fixed point implementation on Pentium Core2Duo (2.13 GHz) and ARM9 (400 MHz) takes 10 seconds and one hour each to process a frame.

• Journal of Broadcast Engineering
• /
• v.19 no.3
• /
• pp.329-341
• /
• 2014

Kernel-based mean-shift object tracking has gained more interests nowadays, with the aid of its feasibility of reliable real-time implementation of object tracking. This algorithm calculates the best mean-shift vector based on the color histogram similarity between target model and target candidate models, where the color histograms are usually produced after uniform color-space quantization for the implementation of real-time tracker. However, when the image of target model has a reduced contrast, such uniform quantization produces the histogram model having large values only for a few histogram bins, resulting in a reduced accuracy of similarity comparison. To solve this problem, a non-uniform quantization algorithm has been proposed, but it is hard to apply to real-time tracking applications due to its high complexity. Therefore, this paper proposes a fast non-uniform color-space quantization method using the histogram equalization, providing an adjusted histogram distribution such that the bins of target model histogram have as many meaningful values as possible. Using the proposed method, the number of bins involved in similarity comparison has been increased, resulting in an enhanced accuracy of the proposed mean-shift tracker. Simulations with various test videos demonstrate the proposed algorithm provides similar or better tracking results to the previous non-uniform quantization scheme with significantly reduced computation complexity.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.25 no.8A
• /
• pp.1085-1095
• /
• 2000

In this paper, we have defined the rules and the algorithm for diagnosis and recovery of routing configuration fault on a system. By using them, we have implemented the Java-based system that can manage routing configuration fault automatically. To manage routing configuration fault, the production rule for network configuration management, the production rule for routing configuration fault diagnosis, and the production rule for routing configuration fault recovery have been proposed. Rule-based routing configuration fault management system has been implemented on the basis of backward chaining algorithm and applied for meta rules for the purpose of interconnecting the production rules. We have derived the experimental result from transition process of the rules, the Blackboard, the goals based on scenarios. Through the implementation of dynamically applicable system in heterogeneous and rapidly changing network environments, we have proposed the methodology for network configuration fault management. Also, we expect that network configuration manager can reduce time and cost wasted for routing configuration fault management.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.10 no.7
• /
• pp.1237-1245
• /
• 2006

This paper presented a hardware implementation of ARIA, which is a Korean standard l28-bit block cryptography algorithm. In this work, ARIA was designed technology-independently for application such as ASIC or core-based designs. ARIA algorithm was fitted in FPGA without additional components of hardware or software. It was confirmed that the rate of resource usage is about 19% in Altera EPXAl0F1020CI and the resulting design operates stably in a clock frequency of 36.35MHz, whose encryption/decryption rate was 310.3Mbps. Consequently, the proposed hardware implementation of ARIA is expected to have a lot of application fields which need high speed process such as electronic commerce, mobile communication, network security and the fields requiring lots of data storing where many users need processing large amount of data simultaneously.