• Journal of Broadcast Engineering
• /
• v.27 no.6
• /
• pp.885-896
• /
• 2022

Motor failure in manufacturing plays an important role in future A/S and reliability. Motor failure is detected by measuring sound, current, and vibration. For the data used in this paper, the sound of the car's side mirror motor gear box was used. Motor sound consists of three classes. Sound data is input to the network model through a conversion process through MelSpectrogram. In this paper, various methods were applied, such as data augmentation to improve the performance of classifying fault motors and various methods according to class imbalance were applied resampling, reweighting adjustment, change of loss function and representation learning and classification into two stages. In addition, the curriculum learning method and self-space learning method were compared through a total of five network models such as Bidirectional LSTM Attention, Convolutional Recurrent Neural Network, Multi-Head Attention, Bidirectional Temporal Convolution Network, and Convolution Neural Network, and the optimal configuration was found for motor sound classification.

• Korean Journal of Applied Biomechanics
• /
• v.18 no.4
• /
• pp.171-177
• /
• 2008

This work intends to investigate the effects of pedaling directions on the muscle actions during the bicycle's uphill propulsion. A test rig was developed that consists of a bicyle with a special planetary geartrain, a height-adjustable treadmill, a rear-wheel support and a magnetic brake. A three-dimensional motion analysis was performed for measuring kinematic characteristics of the forward backward pedaling and the electromygraphy(EMG) measurements were simultaneously performed for estimating the muscle actions of the leg. In this work, four muscles are considered including Gastrocnemius muscle(GM), Vastus lateralis(VL), Tibialis anterior(TA) and Soleus(SOL) while the uphill slope is varied from $0^{\circ}$ to $6^{\circ}$. Raw EMG signals were first processed through the root-mean-square(RMS) averaging and then ensemble curves were derived by averaging the EMG RMS envelopes over 50 consecutive cycles. Results show that both the kinemactic characteristics and the muscle actions are significantly affected by the pedaling direction. The crank speed of the forward pedaling is higher but the difference in speed is reduced as the slope is increased. The ensemble curves of the :ac signals clearly exhibit some differences in their patterns, peak values and the corresponding locations with respect to the crank angle. The peak values of most EMG signals are higher for the forward pedaling regardless of the slope magnitude. However, the averages of the EMG signals are not observed to have a similar relationship with the pedaling direction, which seems to be affected by several factors such as less experience of the participants' backward pedaling. inappropriate bicycle design for the backward pedaling. These limitations will be further considered in future work.

• Proceedings of the KOR-BRONCHOESO Conference
• /
• 1977.06a
• /
• pp.6.1-6
• /
• 1977

Among the various audiometric test batteries, pure tone and speech audiometry was carried as routine. During the past 20 years, there have been works on speech word list for Korean language. As a preliminary work for standardization of Korean speech word list, the authors are intended to study on the factors of test methods with these proposal speech word lists. Now, authors has studied as followings methods and materials and the results were obtained as followings. A. Materials and Methods a. Word List 1. Paik Chun Kee (Paik) 2. Soh Jin Myoung (Soh) 3. Cynn Kyu Shik (Cynn) 4. Lee Jong Dam (Lee) 5. Hahm Tae Young (Hahm) b. Test Method 1. ascending 2. descending c. Test Material 1. live voice 2. Recorded male voice 3. Recorded female voice B. Results 1. Score of ascending method was better result than descending method. 2. Speech reception threshold and discrimination score were better in recorded female voice, live voice and recorded male voice in that order. 3. Speech reception thresholds among tested word lists were Hahm, Lee, Soh, Cynn and Paik in that order. 4. Discrimination score was highest with Cynn and Lee, Soh, Hahm, Paik in that order.