• Journal of Aerospace System Engineering
• /
• v.16 no.4
• /
• pp.45-52
• /
• 2022

In the field of UAM, which is attracting attention as a next-generation transportation system, technology developments for using UAVs have been actively conducted in recent years. Since UAVs adopted with these technologies are mainly operated in urban areas, it is imperative that accidents are prevented. However, it is not easy to predict the abnormal flight state of an UAV causing a crash, because of its strong non-linearity. In this paper, we propose a method for predicting a flight state of an UAV, based on a CNN-LSTM hybrid model. To predict flight state variables at a specific point in the future, the proposed model combines the CNN model extracting temporal and spatial features between flight data, with the LSTM model extracting a short and long-term temporal dependence of the extracted features. Simulation results show that the proposed method has better performance than the prediction methods, which are based on the existing artificial neural network model.

• Transactions of the Korean Society of Automotive Engineers
• /
• v.3 no.1
• /
• pp.33-44
• /
• 1995

As the occupant safety receives more attention from automobile industries. protection systems have been developed quite well. Developed protection systems must be evaluated through real tests in crash environment Since the real tests are extremely expensive. computer simulations are replaced for some prediction of the real test In the computer simulation. it is very crucial to express the real environment precisely in the modeling precess. The energy absorbing(EA) steering system has a very important rote in vehicle crashes because the occupant can hit the system directly. In this study. the EA steering system is modeled precisely. analyzed for the safely and designed by an optimization technology. First. the EA steering system is disassembled by parts and modeled by segments and joints. The segments are modeled by rigid bodies in motion and they have resistances in contact. Spring-damper elements and force-deflection curves are utilized to represent the joints. The body block test is cal lied out to validate. the modeling. When the test results are not enough for the detailed modeling. the differences between tests and simulations are minimized to calculate unknown parameters using optimization. The established model is applied to a crash simulation of a full-car model and tuned again. After the modeling is finished. components of the steering system are designed by an optimization algorithm. In the optimization process. the compound injury of a driver is defined and minimized to determine the chracteristics of the components. The second. order approximation algorithm has been adopted for the optimization.

• Journal of Auto-vehicle Safety Association
• /
• v.15 no.1
• /
• pp.55-62
• /
• 2023

In this study, a chest deflection is predicted by introducing a deep learning technique with the results of the frontal impact of the USNCAP conducted for 110 car models from MY2018 to MY2020. The 120 data are divided into training data and test data, and the training data is divided into training data and validation data to determine the hyperparameters. In this process, the deceleration data of each vehicle is averaged in units of 10 ms from crash pulses measured up to 100 ms. The performance of the deep learning model is measured by the indices of the mean squared error and the mean absolute error on the test data. A DNN (Deep Neural Network) model can give different predictions for the same hyperparameter values at every run. Considering this, the mean and standard deviation of the MSE (Mean Squared Error) and the MAE (Mean Absolute Error) are calculated. In addition, the deep learning model performance according to the inclusion of CVW (Curb Vehicle Weight) is also reviewed.

• Proceedings of the Korean Institute of Building Construction Conference
• /
• 2018.11a
• /
• pp.122-123
• /
• 2018

Although the disaster rate of the industry as a whole is on a downward trend, the disaster rate of the construction industry is on an ongoing trend. Therefore, in this study, we analyzed safety accident statistical data of the construction site over the past three years. As a result of the analysis, the incidence of disasters at small construction sites was very high. And the proportion of disaster occurred for workers who worked in less than 6 months even roughly 92.6%. In addition, as a result of analyzing the form of disaster occurrence, the crash was 34.1% and the fall was 15.1%. The analysis results of these construction safety accidents are to provide as a basic material for developing a policy that can prevent safety accidents and a safety accident prediction model.

• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.20 no.6
• /
• pp.203-213
• /
• 2021

A vehicle crash occurs due to various factors such as the geometry of the road section, traffic, and driver characteristics. A safety performance function has been used in many studies to estimate the relationship between vehicle crash and road factors statistically. And depends on the purpose of the analysis, various characteristic variables have been used. And various characteristic variables have been used in the studies depending on the purpose of analysis. The existing domestic studies generally reflect the average characteristics of the sections by quantifying the traffic volume in macro aggregate units such as the ADT, but this has a limitation that it cannot reflect the real-time changing traffic characteristics. Therefore, the need for research on effective aggregation units that can flexibly reflect the characteristics of the traffic environment arises. In this paper, we develop a safety performance function that can reflect the traffic characteristics in detail with an aggregate unit for one hour in addition to the daily model used in the previous studies. As part of the present study, we also perform a comparison and evaluation between models. The safety performance function for daily and hourly units is developed using a negative binomial regression model with the number of accidents as a dependent variable. In addition, the optimal negative binomial regression model for each of the hourly and daily models was selected, and their prediction performances were compared. The model and evaluation results presented in this paper can be used to determine the risk factors for accidents in the highway section considering the dynamic characteristics. In addition, the model and evaluation results can also be used as the basis for evaluating the availability and transferability of the hourly model.