• IEMEK Journal of Embedded Systems and Applications
• /
• v.18 no.2
• /
• pp.51-58
• /
• 2023

This paper proposes a deep learning method for estimating the heart rate from facial videos. Our proposed method estimates remote photoplethysmography (rPPG) signals to predict the heart rate. Although there have been proposed several methods for estimating rPPG signals, most previous methods can not be utilized in low-power single board computers due to their computational complexity. To address this problem, we construct a lightweight student model and employ a knowledge distillation technique to reduce the performance degradation of a deeper network model. The teacher model consists of 795k parameters, whereas the student model only contains 24k parameters, and therefore, the inference time was reduced with the factor of 10. By distilling the knowledge of the intermediate feature maps of the teacher model, we improved the accuracy of the student model for estimating the heart rate. Experiments were conducted on the UBFC-rPPG dataset to demonstrate the effectiveness of the proposed method. Moreover, we collected our own dataset to verify the accuracy and processing time of the proposed method on a real-world dataset. Experimental results on a NVIDIA Jetson Nano board demonstrate that our proposed method can infer the heart rate in real time with the mean absolute error of 2.5183 bpm.

• Advances in nano research
• /
• v.17 no.1
• /
• pp.19-32
• /
• 2024

Functionally graded-carbon nanotube (FG-CNT) is expected to be a new generation of materials with a wide range of potential applications in technological fields such as aerospace, defense, energy, and structural industries. In this paper, an exact finite strip method for functionally graded-carbon nanotube sandwich plates is developed using first-order shear deformation theory to get the exact natural frequencies of the plates. The face sheets of the plates are made of FG-CNT with continuous and smooth grading based on the power law index. The equations of motion have been generated based on the Hamilton principle. By extracting the exact stiffness matrix for any strip of the sandwich plate as a non-algebraic function of natural frequencies, it is possible to calculate the exact free vibration frequencies. The accuracy and efficiency of the current method is established by comparing its findings to the results of the literature works. Examples are presented to prove the efficiency of the generated method to deal with various problems, such as the influence of the length-to-height ratio, the power law index, and a core-to-face sheet thickness of the single and multi-span sandwich plates with various boundary conditions on the natural frequencies. The exact results obtained from this analysis can check the validity and accuracy of other numerical methods.

• Advances in nano research
• /
• v.17 no.2
• /
• pp.109-124
• /
• 2024

In the era of smart manufacturing, precise prediction of springback-a common issue in ultra-thin sheet metal forming- and forming limits are critical for ensuring high-quality production and minimizing waste. This paper presents a novel approach that leverages the Internet of Things (IoT) and Artificial Neural Networks (ANN) to enhance springback and forming limits prediction accuracy. By integrating IoT-enabled sensors and devices, real-time data on material properties, forming conditions, and environmental factors are collected and transmitted to a central processing unit. This data serves as the input for an ANN model, which is trained with crystal plasticity simulations and experimental data to predict springback with high precision. Our proposed system not only provides continuous monitoring and adaptive learning capabilities but also facilitates real-time decision-making in manufacturing processes. Experimental results demonstrate significant improvements in prediction accuracy compared to traditional methods, highlighting the potential of IoT and ANN integration in advancing smart manufacturing. This approach promises to revolutionize quality control and operational efficiency in the industry, paving the way for more intelligent and responsive manufacturing systems.

• Bulletin of the Korean Chemical Society
• /
• v.32 no.4
• /
• pp.1170-1178
• /
• 2011

In the present study, the approach of high accuracy mass measurements for phospholipid identifications was evaluated using a 7 T ESI-FTMS/linear ion trap MS/MS. Experiments were carried out for porcine brain, bovine liver, and soybean total lipid extracts in both positive and negative ion modes. In total, 59, 55, and 18 phospholipid species were characterized in the positive ion mode for porcine brain, bovine liver, and soybean lipid extracts, respectively. Assigned lipid classes were PC, PE, PEt, PS, and SM. In the negative ion mode, PG, PS, PA, PE, and PI classes were observed. In the negative ion mode, for porcine brain, bovine liver, and soybean lipid extracts, 28, 34, and 29 species were characterized, respectively. Comparison of our results with those obtained by other groups using derivatization-LC-APCI MS and nano-RP-LC-MS/MS showed that our approach can characterize PC species as effectively as those methods could. In conclusion, we demonstrated that high accuracy mass measurements of total lipid extracts using a high resolution FTMS, particularly, 7T FTMS, plus ion-trap MS/MS are very useful in profiling lipid compositions in biological samples.

• Advances in nano research
• /
• v.12 no.5
• /
• pp.515-527
• /
• 2022

The application of fibers in concrete obviously enhances the properties of concrete, also the application of natural fibers in concrete is raising due to the availability, low cost and environmentally friendly. Besides, predicting the mechanical properties of concrete in general and shear strength in particular is highly significant in concrete mixture with fiber nanocomposite reinforced concrete (FRC) in construction projects. Despite numerous studies in shear strength, determining this strength still needs more investigations. In this research, Adaptive Neuro-Fuzzy Inference System (ANFIS) have been employed to determine the strength of reinforced concrete with fiber. 180 empirical data were gathered from reliable literature to develop the methods. Models were developed, validated and their statistical results were compared through the root mean squared error (RMSE), determination coefficient (R²), mean absolute error (MAE) and Pearson correlation coefficient (r). Comparing the RMSE of PSO (0.8859) and ANFIS (0.6047) have emphasized the significant role of structural parameters on the shear strength of concrete, also effective depth, web width, and a clear depth rate are essential parameters in modeling the shear capacity of FRC. Considering the accuracy of our models in determining the shear strength of FRC, the outcomes have shown that the R² values of PSO (0.7487) was better than ANFIS (2.4048). Thus, in this research, PSO has demonstrated better performance than ANFIS in predicting the shear strength of FRC in case of accuracy and the least error ratio. Thus, PSO could be applied as a proper tool to maximum accuracy predict the shear strength of FRC.

• Proceedings of the KWS Conference
• /
• 2009.11a
• /
• pp.29-29
• /
• 2009

On this research, laser welding technology for manufacturing automobile body is studied. Laser welding technology is one of the important technologies used in the manufacturing of lighter, safer automotive bodies at a high level of productivity; the leading automotive manufacturers have replaced spot welding with laser welding in the process of car body assembly. Korean auto manufacturers are developing and applying the laser welding technology using a high output power Nd:YAG laser and a 6-axes industrial robot. On the other hand, the robot-based remote laser welding system was equipped with a long focal laser scanner system in robotic end effect. Laser system, robot system, and scanner system are used for realizing the high speed laser welding system. The remote laser welding system and industrial robotic system are used to consist of robot-based remote laser welding system. The robot-based remote laser welding system is flexible and able to improve laser welding speed compared with traditional welding as spot welding and laser welding. The robot-based remote laser systems used in this study were Trumpf's 4kW Nd:YAG laser (HL4006D) and IPG's 1.6kW Fiber laser (YLR-1600), while the robot systems were of ABB's IRB6400R (payload:120kg) and Hyundai Heavy Industry's HX130-02 (payload:130kg). In addition, a study of quality evaluation and monitoring technology for the remote laser welding was conducted. The welding joints of steel plate and steel plate coated with zinc were butt and lapped joints. The quality testing of the laser welding was conducted by observing the shape of the beads on the plate and the cross-section of the welded parts, analyzing the results of mechanical tension test, and monitoring the plasma intensity and temperature by using UV and IR detectors. Over the past years, Trumf's 4kW Nd:YAG laser and ABB's IRB6400R robot system was used. Nowadays, the new laser source, robot and laser scanner system are used to increase the processing speed and to improve the efficiency of processes. This paper proposes the robot-based remote laser welding system as a means of resolving the limited welding speed and accuracy of conventional laser welding systems.

• JSTS:Journal of Semiconductor Technology and Science
• /
• v.4 no.4
• /
• pp.269-274
• /
• 2004

Current delivery system, in which the analog current produced by a unique DAC circuit is stored into a current-memory circuit and delivered in a time-divided sequence, shows variation of output current as low as 4% in a current source data-driver IC for AM-OLED driven by a current-programmed method without any fuse repairing after fabrication. This driver IC has 54 outputs and can sink constant current as low as 3 ${\mu}A$ with 6-bit analog levels. Such a low current level without variation can hardly be obtained by an ordinary MOS transistor because the current level is in the sub-threshold region and changes exponentially with threshold voltage variation. Thus we adopted a current mirror circuit composed of bipolar transistors to supply well-controlled current within a nano-ampere range.

• Journal of the Korean Society of Manufacturing Technology Engineers
• /
• v.22 no.5
• /
• pp.831-836
• /
• 2013

This study demonstrates the process of fabricating patterns using tribonanolithography (TNL),with laboratory-made micro polycrystalline diamond (PCD) tools that are attached to an atomic force microscope (AFM). The various patterns are easily fabricated using mechanical scratching, under various normal loads, using the PCD tool on single crystal silicon, which is the master mold for replication in this study. Then, polydimethylsiloxane (PDMS) replica molds are fabricated using precise pattern transfer processes. The transferred patterns show high dimensional accuracy as compared with those of TNL-processed silicon micro molds. TNL can reduce the need for high cost and complicated apparatuses required for conventional lithography methods. TNL shows great potential in that it allows for the rapid fabrication of duplicated patterns through simple mechanical micromachining on brittle sample surfaces.

• Journal of the Korean Society of Manufacturing Process Engineers
• /
• v.5 no.1
• /
• pp.45-50
• /
• 2006

This paper shows a new precision actuator of MUTM(miniature universal testing machine) for the testing of compression and tensile load on the MEMS materials and structures. The MUTM consists of a sample holder, an ultraprecision precision actuator(tranlation stage) and load sensor. The precision actuator has been developed for generating displacements with nanometer accuracy and a dynamic range of 1mm simultaneously. In this paper, it can be made by using the displacement property of bi-morph type PZT, which is able to extend the long range(stroke) according to cantilever size. However, it is not enough to be generated for compression and tensile load in miniature universal testing machine. Therefore, three dozen bi-morph type PZTs are used for generating the load. The load and displacement of the precision actuator are 35g and 0.4mm respectively.

• Advances in nano research
• /
• v.6 no.2
• /
• pp.93-112
• /
• 2018

An analytical solution of the buckling governing equations of functionally graded piezoelectric (FGP) nanobeams obtained by using a developed third-order shear deformation theory is presented. Electro-mechanical properties of FGP nanobeam are supposed to change continuously in the thickness direction based on power-law model. To capture the small size effects, Eringen's nonlocal elasticity theory is adopted. Employing Hamilton's principle, the nonlocal governing equations of a FG nanobeams made of piezoelectric materials are obtained and they are solved using Navier-type analytical solution. Results are provided to show the effect of different external electric voltage, power-law index, nonlocal parameter and slenderness ratio on the buckling loads of the size-dependent FGP nanobeams. The accuracy of the present model is verified by comparing it with nonlocal Timoshenko FG beams. So, this study makes the first attempt for analyzing buckling behavior of higher order shear deformable FGP nanobeams.