• ETRI Journal
• /
• v.44 no.6
• /
• pp.1004-1019
• /
• 2022

For autonomous driving without high-definition maps, we present a model capable of generating multiple plausible paths from egocentric images for autonomous vehicles. Our generative model comprises two neural networks: feature extraction network (FEN) and path generation network (PGN). The FEN extracts meaningful features from an egocentric image, whereas the PGN generates multiple paths from the features, given a driving intention and speed. To ensure that the paths generated are plausible and consistent with the intention, we introduce an attentive discriminator and train it with the PGN under a generative adversarial network framework. Furthermore, we devise an interaction model between the positions in the paths and the intentions hidden in the positions and design a novel PGN architecture that reflects the interaction model for improving the accuracy and diversity of the generated paths. Finally, we introduce ETRIDriving, a dataset for autonomous driving, in which the recorded sensor data are labeled with discrete high-level driving actions, and demonstrate the state-of-the-art performance of the proposed model on ETRIDriving in terms of accuracy and diversity.

• Korean Journal of Construction Engineering and Management
• /
• v.24 no.5
• /
• pp.73-82
• /
• 2023

Optimization of Construction Site Layout Planning (CSLP) heavily relies on workers' travel paths. However, traditional path generation approaches predominantly focus on the shortest path, often neglecting critical variables such as individual wayfinding tendencies, the spatial arrangement of site objects, and potential hazards. These oversights can lead to compromised path simulations, resulting in less reliable site layout plans. While Deep Reinforcement Learning (DRL) has been proposed as a potential alternative to address these issues, it has shown limitations. Despite presenting more realistic travel paths by considering these variables, DRL often struggles with efficiency in complex environments, leading to extended learning times and potential failures. To overcome these challenges, this study introduces a refined model that enhances spatial navigation capabilities and learning performance by integrating workers' visibility into the reward functions. The proposed model demonstrated a 12.47% increase in the pathfinding success rate and notable improvements in the other two performance measures compared to the existing DRL framework. The adoption of this model could greatly enhance the reliability of the results, ultimately improving site operational efficiency and safety management such as by reducing site congestion and accidents. Future research could expand this study by simulating travel paths in dynamic, multi-agent environments that represent different stages of construction.

• Advances in nano research
• /
• v.16 no.6
• /
• pp.531-548
• /
• 2024

The study focuses on using remote sensing to gather data about the Earth's surface, particularly in urban environments, using satellites and aircraft-mounted sensors. It aims to develop a classification framework for road targets using multi-spectral imagery. By integrating Convolutional Neural Networks (CNNs) with XGBoost, the study seeks to enhance the accuracy and efficiency of road target identification, aiding urban infrastructure management and transportation planning. A novel aspect of the research is the incorporation of quantum sensors, which improve the resolution and sensitivity of the data. The model achieved high predictive accuracy with an MSE of 0.025, R-squared of 0.85, RMSE of 0.158, and MAE of 0.12. The CNN model showed excellent performance in road detection with 92% accuracy, 88% precision, 90% recall, and an f1-score of 89%. These results demonstrate the model's robustness and applicability in real-world urban planning scenarios, further enhanced by data augmentation and early stopping techniques.

• Korean Journal of Construction Engineering and Management
• /
• v.18 no.6
• /
• pp.26-37
• /
• 2017

Preliminary scheduling at the initial stage of planning phase is usually performed with limited information and details. Therefore, the reliability and accuracy of preliminary scheduling is affected by personal experiences and skills of the schedule planners, and it requires enormous managerial effort (or workload). Reusing of historical data of the similar projects is important for efficient preliminary scheduling. However, understanding the structure of historical data and applying them to a new project requires a great deal of experience and knowledge. In this context, this paper propose a framework and methodology for automated preliminary schedule generation based on historical database. The proposed methodology and framework enables to automatically generate CPM schedules for airport projects in the early planning stage in order to enhance the reliability and to reduce the workload by using structured knowledge and experience.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.21 no.12
• /
• pp.200-210
• /
• 2020

The maintenance of educational facilities was assembled in 12 project classifications of the Educational Improvement Program. The priorities were decided by the evaluation scores derived from the condition investigation, and maintenance works were budgeted in the order of priorities. These priorities were a schedule for conducting maintenance and an important criterion for obtaining a construction order. Several restrictions in the condition investigation exist, which derives budgets and conducts maintenance separately based on the priorities. An educational facility manager has a restriction in quantity take-off, which results in an incorrect budget. Discomfort would occur in an educational environment, and a period of infringing safety would increase. This study proposes applying a BIM in the condition investigation and the planning framework for work-type packaging. A BIM supports a budget calculation and derives evaluation scores by linking a repair and an inspection result. The work-type packaging algorithm divides a budget allocation range and derives the result of a grouped work-types applied in an equivalent space and element. As a result of applying cases, it could shorten the duration by approximately 37.4%. Its usability in selecting a grouped work-type was evaluated through an assessment with workers.

• Journal of Intelligence and Information Systems
• /
• v.27 no.3
• /
• pp.199-230
• /
• 2021

Innovative new products and services are being launched through the convergence between heterogeneous industries, and social interest and investment in convergence industries such as AI, big data-based future cars, and robots are continuously increasing. However, in the process of commercialization of convergence new products and services, there are many cases where they do not conform to the existing regulatory and legal system, which causes many difficulties in companies launching their products and services into the market. In response to these industrial changes, the current government is promoting the improvement of existing regulatory mechanisms applied to the relevant industry along with the expansion of investment in new industries. This study, in these convergence industry trends, aimed to analysis the existing regulatory system that is an obstacle to market entry of innovative new products and services in order to preemptively predict regulatory issues that will arise in emerging industries. In addition, it was intended to establish a regulatory impact analysis system to evaluate adequacy and prepare improvement measures. The flow of this study is divided into three parts. In the first part, previous studies on regulatory impact analysis and evaluation systems are investigated. This was used as basic data for the development direction of the regulatory impact framework, indicators and items. In the second regulatory impact analysis framework development part, indicators and items are developed based on the previously investigated data, and these are applied to each stage of the framework. In the last part, a case study was presented to solve the regulatory issues faced by actual companies by applying the developed regulatory impact analysis framework. The case study included the autonomous/electric vehicle industry and the Internet of Things (IoT) industry, because it is one of the emerging industries that the Korean government is most interested in recently, and is judged to be most relevant to the realization of an intelligent information society. Specifically, the regulatory impact analysis framework proposed in this study consists of a total of five steps. The first step is to identify the industrial size of the target products and services, related policies, and regulatory issues. In the second stage, regulatory issues are discovered through review of regulatory improvement items for each stage of commercialization (planning, production, commercialization). In the next step, factors related to regulatory compliance costs are derived and costs incurred for existing regulatory compliance are calculated. In the fourth stage, an alternative is prepared by gathering opinions of the relevant industry and experts in the field, and the necessity, validity, and adequacy of the alternative are reviewed. Finally, in the final stage, the adopted alternatives are formulated so that they can be applied to the legislation, and the alternatives are reviewed by legal experts. The implications of this study are summarized as follows. From a theoretical point of view, it is meaningful in that it clearly presents a series of procedures for regulatory impact analysis as a framework. Although previous studies mainly discussed the importance and necessity of regulatory impact analysis, this study presented a systematic framework in consideration of the various factors required for regulatory impact analysis suggested by prior studies. From a practical point of view, this study has significance in that it was applied to actual regulatory issues based on the regulatory impact analysis framework proposed above. The results of this study show that proposals related to regulatory issues were submitted to government departments and finally the current law was revised, suggesting that the framework proposed in this study can be an effective way to resolve regulatory issues. It is expected that the regulatory impact analysis framework proposed in this study will be a meaningful guideline for technology policy researchers and policy makers in the future.

• Journal of Digital Convergence
• /
• v.13 no.10
• /
• pp.225-234
• /
• 2015

This study focused on building the convergence framework of decision-making for camping site selection attributes. The first stage of study is to identify the camping motivations. The camping motivations are composed of 5 factors. The second stage is to identify the camping site selection attributes. They are composed of 4 factors. The relationships between camping site selection attributes and camping motivations are examined through the multi-regression analysis. The camping site selection attributes were significantly influenced by thr type of camping motivation. Lastly, the relationship between camping site selection attributes and camper's satisfaction and behavioral intention are testified. The implications are that camping site operators and developers need to understand the segmented camping motivations and site selection attributes. They could be the guideline of implementing the developing planning strategies of camping site.

• Korean Journal of Construction Engineering and Management
• /
• v.13 no.3
• /
• pp.78-88
• /
• 2012

Previous construction failure researches were focused on the utilization plan based on failure information and development of failure classification. However, it has limitation to set up the plan for prevention of construction failure due to the lack of the number of on-site staffs. In order to prevent effectively construction failure, a prevention plan should be established through quantitative evaluation of failure causes. The purpose of this study is to suggest the assessment method for selection Critical Management Factor(CMF) and to analyze the CMF on the apartment structural framework using FMEA(Failure Mode and Effective Analysis) which is one of the methods of quantitative evaluation. The element of risk evaluation separated degree of failure risk and prevention respectively. The assessment method for selection of CMF can be utilized for planning proactive solutions on the failure, and it can be also selected critical factors about each project phases, type of facility and construction work.

• Proceedings of the Korean Institute Of Construction Engineering and Management
• /
• autumn
• /
• pp.509-512
• /
• 2003

The Critical Path Method(CPM) is an effective tool used for planning and scheduling. One of strong point in the CPM is what can calculate float. Float is able to prolong without having an effect on overall schedule of project, however concept and scope about ownership is not definite, because it is a by-product of project. Thus participants have had many dispute in using float because of their interests. In recent years, a few theories have evolved in an attempt to solve this problem. But the prior research did not make a reasonable distribution of float because their criteria of distribution was based on ratio for total project time. Actually, an application of float is achieved by participant's delay risk. Therefore. this paper proposes criteria for float distribution using delay risks, and the framework for the assessment of the delay risks.

• Journal of the Korean Society for information Management
• /
• v.35 no.3
• /
• pp.271-286
• /
• 2018

The purpose of this study is to measure the outcomes of a program provided by a public library using the evaluation framework based on Logic Model. A reading program for children which was operated by a public library in Seoul was selected. The outcome evaluation was started with the analysis of the reading program process including planning, operation, and evaluation. Based on the analysis, a logic model framework for outcome evaluation was developed. For evaluation, user, bibliography, and circulation data were collected from the library KOLAS system. Additionally, the participant information were extracted from the final report drafted after the program. The research results show that the number of circulation of program participants was increased after the program. In addition, the range of reading topic was expanded. These findings indicate that the reading program is an effective program for promoting children's reading habit and that outcome evaluation might be a valid tool to measure the effectiveness of public library programs.