• Proceedings of the Korean Society of Precision Engineering Conference
• /
• 2002.10a
• /
• pp.530-535
• /
• 2002

Pneumatic control system has been applied to build many industrial automation systems. But most of them are sequence control type because of their low costs, safety, reliability, etc. Pneumatic servo system is rarely applied to real industrial fields because accurate position control is very difficult due to its nonlinearity and compressibility of air. In pneumatic servo control system, a pneumatic servo valve can be applied, But it is very expensive and has no advantage of low cost compared with a common pneumatic system. This paper is concerned with the accurate position control of a rodless pneumatic cylinder using on/off solenoid valve. A novel Intelligent Modified Pulse Width Modulation(MPWM) is newly proposed. The control performance of this pneumatic cylinder depends on the external loads. To overcome this problem, switching of control parameter using artificial neural network is newly proposed, which estimates external loads on rodless pneumatic cylinder using this training neural network. As an underlying controller, a state feedback controller using position, velocity and acceleration is applied in the switching control the system. The effectiveness of the proposed control algorithms are demonstrated through experiments nth various loads.

• Journal of the Korea Society of Computer and Information
• /
• v.29 no.7
• /
• pp.11-20
• /
• 2024

In this paper, we propose the BAG (Bit-width Aware Generator) and the Intermediate Layer Knowledge Distillation using Channel-wise Attention to reduce the knowledge gap between a quantized network, a full-precision network, and a generator in GDFQ (Generative Data-Free Quantization). Since the generator in GDFQ is only trained by the feedback from the full-precision network, the gap resulting in decreased capability due to low bit-width of the quantized network has no effect on training the generator. To alleviate this problem, BAG is quantized with same bit-width of the quantized network, and it can generate synthetic images, which are effectively used for training the quantized network. Typically, the knowledge gap between the quantized network and the full-precision network is also important. To resolve this, we compute channel-wise attention of outputs of convolutional layers, and minimize the loss function as the distance of them. As the result, the quantized network can learn which channels to focus on more from mimicking the full-precision network. To prove the efficiency of proposed methods, we quantize the network trained on CIFAR-100 with 3 bit-width weights and activations, and train it and the generator with our method. As the result, we achieve 56.14% Top-1 Accuracy and increase 3.4% higher accuracy compared to our baseline AdaDFQ.

• Nuclear Engineering and Technology
• /
• v.54 no.3
• /
• pp.834-841
• /
• 2022

System thermal-hydraulic (STH) code is adopted for nuclear safety analysis. The critical flow model (CFM) is significant for the accuracy of STH simulation. To overcome the defects of current CFMs (low precision or long calculation time), a CFM based on a genetic neural network (GNN) has been developed in this work. To build a powerful model, besides the critical mass flux, the critical pressure and critical quality were also considered in this model, which was seldom considered before. Comparing with the traditional homogeneous equilibrium model (HEM) and the Moody model, the GNN model can predict the critical mass flux with a higher accuracy (approximately 80% of results are within the ±20% error limit); comparing with the Leung model and the Shannak model for critical pressure prediction, the GNN model achieved the best results (more than 80% prediction results within the ±20% error limit). For the critical quality, similar precision is achieved. The GNN-based CFM in this work is meaningful for the STH code CFM development.

• Korean Journal of Radiology
• /
• v.25 no.8
• /
• pp.749-755
• /
• 2024

The recent surge in the incidence of small papillary thyroid cancers (PTCs) has been linked to the widespread use of ultrasonography, thereby prompting concerns regarding overdiagnosis. Active surveillance (AS) has emerged as a less invasive alternative management strategy for low-risk PTCs, especially for PTCs measuring ≤1 cm in maximal diameter. Recent studies report low disease progression rates of low-risk PTCs ≤1 cm under AS. Ongoing research is currently exploring the feasibility of AS for larger PTCs (<20 mm). AS protocols include meticulous ultrasound assessment, emphasis on standardized techniques, and a multidisciplinary approach; they involve monitoring the nodules for size, growth, potential extrathyroidal extension, proximity to the trachea and recurrent laryngeal nerve, and potential cervical nodal metastases. The criteria for progression, often defined as an increase in the maximum diameter of the PTC, warrant a review of precision and ongoing examinations. Challenges exist regarding the reliability of volume measurements for defining PTC disease progression. Although ultrasonography plays a pivotal role, challenges in assessing progression and minor extrathyroidal extension underscore the importance of a multidisciplinary approach in disease management. This comprehensive overview highlights the evolving landscape of AS for PTCs, emphasizing the need for standardized protocols, meticulous assessments, and ongoing research to inform decision-making.

• Journal of the Korean Society of Manufacturing Process Engineers
• /
• v.6 no.1
• /
• pp.27-33
• /
• 2007

Hard turning on modern lathes becomes a realistic replacement for many grinding applications. Because CBN tools are expensive, excessive tool wear can eliminate economic advantages of hard turning. This paper describes a study of investigating the cutting force and the characteristics of tool wear in hard turning of hardened steels, AISI 52100. Cutting forces generated using CBN tools have been evaluated. The radial thrust cutting force was the largest among three cutting force components. It increased dramatically as a result of progressive tool wear. On the other hand, the result shows significantly different wear characteristics between high CBN and low CBN. Backpropagation neural network was used for the estimation of tool wear. The networks were achieved the reliability of 96.3% even when the spindle speed and feed rate are changed.

• Proceedings of the Korean Society of Precision Engineering Conference
• /
• 2005.10a
• /
• pp.769-772
• /
• 2005

This paper deals with the aspects of die design for the multistage fine tooth hub gear in the cold forging process. In order to manufacture the cold forged product fur the precision hub gear used as the ARD 370 system of bicycle, it examines the influences of different designs on the metal flow through experiments and FE-simulation. To find the combination of design parameters which minimize the damage value, the low gear length, upper gear length and inner diameter as design parameters are considered. An orthogonal fraction factorial experiment is employed to study the influence of each parameter on the objective function or characteristics. The optimal punch shape of fine tooth hub gear is designed using the results of FE-simulation and the artificial neural network. To verify the optimal punch shape, the experiments of the cold forging of the hub gear are executed.

• Journal of the Korean Society for Precision Engineering
• /
• v.26 no.4
• /
• pp.82-90
• /
• 2009

Pneumatic cylinder is one of the low cost actuation sources which have been applied in industrial and prosthetic application since it has a high power/weight ratio, a high-tension force and a long durability However, the control problems of pneumatic systems, oscillatory motion and compliance, have prevented their widespread use in advanced robotics. To overcome these shortcomings, a number of newer pneumatic actuators have been developed such as McKibben Muscle, Rubber Actuator and Pneumatic Artificial Muscle (PAM) Manipulators. In this paper, one solution for position control of a robot arm, which is driven by two pneumatic artificial muscles, is presented. However, some limitations still exist, such as a deterioration of the performance of transient response due to the changes in the external load of the robot arm. To overcome this problem, a switching algorithm of the control parameter using a learning vector quantization neural network (LVQNN) is proposed in this paper. This estimates the external load of the pneumatic artificial muscle manipulator. The effectiveness of the proposed control algorithm is demonstrated through experiments with different external working loads.

• Journal of the Korean Society for Precision Engineering
• /
• v.11 no.1
• /
• pp.138-149
• /
• 1994

This paper introduces a new monitoring technique which utilizes an adaptive signal processing for feature generation, coupled with a multilayered merual network for pattern recognition. The cutting force signal in face milling operation was modeled by a low order discrete autoregressive model, shere parameters were estimated recursively at each sampling instant using a parameter adaptation algorithm based on an RLS(recursive least square) method with discounted measurements. The influences of the adaptation algorithm parameters as well as some considerations for modeling on the estimation results are discussed. The sensitivity of the extimated model parameters to the tool state(new and worn tool)is presented, and the application of a multilayered neural network to tool state monitoring using the previously generated features is also demonstrated with a high success rate. The methodology turned out to be quite suitable for in-process tool wear monitoring in the sense that the model parameters are effective as tool state features in milling operation and that the classifier successfully maps the sensors data to correct output decision.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.4 no.3
• /
• pp.243-257
• /
• 2010

Rapid developments in wireless sensor networks have extended many applications, hence, many studies have developed wireless sensor network positioning systems for indoor environments. Among those systems, the Global Position System (GPS) is unsuitable for indoor environments due to Line-Of-Sight (LOS) limitations, while the wireless sensor network is more suitable, given its advantages of low cost, easy installation, and low energy consumption. Due to the complex settings of indoor environments and the high demands for precision, the implementation of an indoor positioning system is difficult to construct. This study adopts a low-cost positioning method that does not require additional hardware, and uses the received signal strength (RSS) values from the receiver node to estimate the distance between the test objects. Since many objects in indoor environments would attenuate the radio signals and cause errors in estimation distances, knowing the path loss exponent (PLE) in an environment is crucial. However, most studies preset a fixed PLE, and then substitute it into a radio propagation loss model to estimate the distance between the test points; such method would lead to serious errors. To address this problem, this study proposes a Path Loss Exponent Estimation Algorithm, which uses only four beacon nodes to construct a radio propagation loss model for an indoor environment, and is able to provide enhanced positioning precision, accurate positioning services, low cost, and high efficiency.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.18 no.5
• /
• pp.147-153
• /
• 2018

Beacon enables one to measure distance indoors based on low-power Bluetooth low energy (BLE) technology, while GPS (Global Positioning System) only can be used outdoors. In measuring indoor distance using Beacon, RSSI (Received Signal Strength Indication) is considered as the one of the key factors, however, it is influenced by various environmental factors so that it causes the huge gap between the estimated distance and the real. In order to handle this issue, we propose the adaptive ANN (Artificial Neural Network) based approach to measuring the exact distance using Beacon. First, we has carried out the preprocessing of the RSSI signals by applying the extended Kalman filter and the signal stabilization filter into decreasing the noise. Then, we suggest the multi-layered ANNs, each of which layer is learned by specific training data sets. The results showed an average error of 0.67m, a precision of 0.78.