Environmentally Sustainable Management of 3D Printing Network: Decision Support for 3D Printing Work Allocation

Jungmok Ma 한국정밀공학회 2020 International Journal of Precision Engineering and Vol.21 No.3

The purpose of this study is to provide a model for environmentally sustainable management of 3D printing network systems. The proposed model provides not only a flexible structure to describe 3D printing processes but also a computational structure to find the optimal work allocation plan for minimizing environmental impact. A mathematical model is formulated to assist the optimal part-to-printer allocation decision in 3D printing network systems even under uncertainty. Numerical examples show that the proposed model can determine the operation of shared 3D printers in order to have minimum environmental impact. The proposed model can also deal with data uncertainty and provide robust solutions.

Optimal Product Design for Life Cycle Assessment (LCA) with the Case Study of Universal Motors

Ma, Jungmok,Kim, Namhun Korean Society for Precision Engineering 2016 International Journal of Precision Engineering and Vol.17 No.9

Though life cycle assessment can provide the analytic basis of environmental impact of products for engineering designers, it requires not only a significant amount of time but also a detailed bill of materials of the products to determine the optimal product designs. To overcome the barrier between life cycle assessment and design, in this paper, the optimal product design for life cycle assessment is studied. Instead of comparing a few design alternatives due to the barrier, design variables are allowed to vary as an optimization problem to minimize environmental impact. As a numerical example, the design of universal electric motors was demonstrated. Under the five different design cases, the optimal designs could be achieved to minimize the total environmental impact of universal motors.

Robust optimal usage modeling of product systems for environmental sustainability

Jungmok Ma 한국CDE학회 2019 Journal of computational design and engineering Vol.6 No.3

Proper modeling of the usage phase in Life Cycle Assessment (LCA) is not only critical due to its high impact among life cycle phases but also challenging due to high variations and uncertainty. Furthermore, when multiple products can be utilized, the optimal product usage should be considered together. The robust optimal usage modeling is proposed in this paper as the framework of usage mod-eling for LCA with consideration of the uncertainty and optimal usage. The proposed method seeks to optimal product usage in order to minimize the environmental impact of the usage phase under uncer-tainty. Numerical examples demonstrate the application of the robust optimal usage modeling and the difference from the previous approaches.

Computational Framework for Usage Stage Modeling of Machines in Life Cycle Assessment

마정목(Jungmok Ma) Korean Society for Precision Engineering 2019 한국정밀공학회지 Vol.36 No.11

Despite the importance of the usage stage in life cycle assessment (LCA), there is a lack of comprehensive studies on the usage stage modeling. Based on the literature review, this paper establishes a general framework of the usage stage modeling by redefining existing models and proposing new models. The proposed computational framework can provide the overview of the current research as well as lead researchers and practitioners to consider proper modeling techniques. The framework includes the representative usage scenario method, usage context modeling, and time series usage modeling. Also, future research directions are suggested with the proposed computational framework.

무기체계 환경성 설계에 관한 기초연구

마정목 ( Ma Jungmok ) 국방대학교 안보문제연구소 2016 국방연구 Vol.59 No.4

현대에 접어들면서 지구온난화, 기후변화, 잠재적인 자원고갈과 환경파괴 및 이에 따른 부수적인 인명 피해까지 환경문제는 지속적으로 인류를 위협하는 문제로 대두되고 있다. 제한된 예산하에 고품질의 무기체계를 적시에 획득하기도 쉽지 않은 국방 획득 환경에서 환경성에 대한 고려는 거의 이루어지지 않는 실정이지만 그럼에도 불구하고 무기체계 환경성에 대해 미리 연구하고 준비해야 하는 이유는 다음과 같다. 무기체계의 장기간 운용 간 주변 환경 및 사용자에게 치명적인 영향을 미칠 수 있고 이는 군 내부는 물론이고 사회적 비용을 발생시켜 전체 수명주기 비용의 증가를 초래한다. 또한 앞으로 생길 수 있는 국제적 환경 규제들에 미리 대처하여야만 국방 분야의 수출 시장을 확대할 수 있으며, 무엇보다 환경성 평가가 장기적인 계획과 준비가 필수적인 분야이기 때문이다. 따라서 본 연구에서는 전과정평가(LCA)라는 환경성 분석 도구를 수학적 모델과 함께 소개하고, 이에 기반을 둔 미국의 동향을 살펴보았다. 현재 미국은 전략적인 성과 계획과 더불어 국방획득 과정에 환경성을 포함할 수 있도록 지속가능성 분석(sustainability analysis)이라는 방안을 제시하고 있다. 본 연구는 대안 분석적 목적이 강한 현 발전추세에서 한 걸음 더 나아가 무기체계 획득 간 환경성 목표를 설정하고 설계 단계에서 이를 관리하며 시험평가 시이를 검증함으로써 전 수명주기 비용이 최소가 되도록 하는 무기체계 획득 방안을 제시하였다. 제시된 환경성 설계 절차와 갖추어야할 입력 자료들을 미리 준비하여 변화되는 환경에서도 고품질의 무기체계를 획득하고 수명주기 비용을 감소시킬 수 있기를 기대해본다. In modern times, our societies are facing critical environmental problems due to global warming, climate change, potential resource depletion, and destruction of the environment, which can cause loss of lives. Since it is very difficult to acquire high quality weapons on time and within tight budget, environmental impacts of weapon systems cannot be a serious concern in the process of their acquisition. Nevertheless, there are some reasons to prepare the research on the management of weapon systems` environmental impacts in advance. While running weapon systems for a long time, it can cause serious damage to the users and environment, which ultimately increases the life cycle cost of weapon systems. Moreover, to export more weapons, required environmental criteria in the future should be satisfied. Lastly, it takes a long time to prepare the implementation of environmental impact assessment. In order to consider the environment and its related costs, this study reviews the concepts and tools of life cycle assessment (LCA) with its mathematical model. The united states developed strategic plans as a guide and sustainability analysis as a tool based on LCA. This study proposes the concept of managing weapon systems` environmental impacts by setting an environmental objective, monitoring environmental impacts, and testing an environmental performance, which can reveal the hidden life cycle costs. It is expecting to minimize the total life cycle cost with the proposed framework.

전투원-로봇 협업시스템 연구

마정목(Jungmok Ma) 육군사관학교 화랑대연구소 2018 한국군사학논집 Vol.74 No.3

While autonomous systems are actively researched with artificial intelligence technologies in defense, humans can be ignored though they are still critical elements in the system. Humans can determine the overall performance and life cycle cost of the autonomous systems if humans are part of the systems. This paper emphasizes that collaboration capability allows the acquired weapon systems to integrate humans and autonomous agents for boosting performance and minimizing life cycle cost. In order to understand the concept and importance of the proposed warfighter-robot collaborative systems, literature reviews are presented in the domain of human factors, robotics, and defense, and two recent studies are analyzed. Based on these, the future research directions of warfighter-robot collaborative systems are provided in the four areas. It is hoped that this study will encourage the follow-up research on the collaborative systems.

미 국방부 무기체계 자율성 훈령 개정에 따른 자율무기체계 정책 분석과 이해

마정목 ( Jungmok Ma ) 국방대학교 안보문제연구소 2023 국방연구 Vol.66 No.2

미국은 국방부 훈령 3000.09를 통해 무기체계의 자율성에 대한 정책을 공개한 최초의 나라이다. 약 10년이 지난 2023년 1월 25일, 미 국방부는 해당 훈령의 개정을 공표하고, 정책의 공개를 통해 글로벌 AI 거버넌스에서 지속 선도적 역할을 할 의지를 이어가고 있다. 그러나 자율무기체계에 대한 정책은 미국 내부에서도 이해가 어려운 주제로 알려져 있다. 본 연구에서는 개정된 미 국방부 훈령을 바탕으로 미국의 자율무기체계 정책에 대한 체계적인 분석을 제공함으로써 관련 정책의 이해도를 높이는데 그 목적이 있다. 이를 위해 ① 자율무기체계와 관련하여 전 세계를 선도하고 있는 미국의 정책은 무엇인지, ② 2012년 이전판과 어떤 차이점이 있는지, ③ 정책적 내용 중 앞으로 개선이 필요한 사항과 추가 해석이 필요한 사항에 대한 검토와 분석을 시행하고자 한다. The United States was the first country to publicly open its policy on autonomy in weapon systems, DoD Directive 3000.09. After about 10 years, on January 25, 2023, U.S. DoD announced the update of the directive. This paper aims at providing systematic analysis of U.S. policy on Autonomous Weapon Systems(AWS) based on DoDD 3000.09 and improving the understanding. There was so much misunderstanding of U.S. policy on AWS not only from the outside of U.S. but also from U.S. DoD personnel. This paper provides 1) the difference between 2012 and 2023 DoDD 3000.09 and 2) the explanation of important policies and concepts based on the understanding of various issues on AWS and AI in defense.

Affordance-based modeling of a humanrobot cooperative system for area exploration

Jeongsik Kim,Jungmok Ma,Namhun Kim 대한기계학회 2020 JOURNAL OF MECHANICAL SCIENCE AND TECHNOLOGY Vol.34 No.2

The cooperation of humans and robots is ubiquitous in modern systems, while human ability to cooperate has been limitedly investigated in terms of systems theory. A formal model is proposed to describe the human capability for the cooperation based on the finite state automata (FSA) and the affordances theory. Unlike the previous studies focused on conceptual approaches, real and virtual experiments are conducted to investigate human actions in a cooperative system with a human and a robot. A modeling scheme is provided to implement agent-based simulations for the cooperative system using the proposed affordance-based FSA. The developed simulation for the cooperation problem can reproduce the patterns of the actual experiments in terms of affordances and supervisory capability. The modular architecture of the agent-based framework allows establishing open-ended algorithms for action selections with their isolated effects under physical constraints, which need to be revised to deal with human-involved cooperation problems.

딥러닝 기반 객체탐지 및 추적기법을 활용한 군 제대규모 판단

김민우(Minwoo Kim),마정목(Jungmok Ma) 한국산학기술학회 2023 한국산학기술학회논문지 Vol.24 No.5

4차 산업혁명 기술의 발전과 인구절벽 등 여러 요인으로 인해 군이 맞닥뜨리게 될 미래 작전환경에서 전투행위의 주체는 무인 중심으로 전환될 것이다. 그중 감시정찰 분야는 가장 먼저 무인체계에 의해 임무가 대체 되어가고 있다. 본 논문에서는 딥러닝 기반의 알고리즘을 활용하여 지상 감시 간 기동하는 군 제대의 규모를 판단하기 위한 연구를 수행하였다. 딥러닝 알고리즘 중 객체탐지와 객체추적 기법을 활용하여 영상 내 특정 무기체계를 식별하고 탐지된 객체의 계수를 통해 제대 규모를 판단하였다. 특히 기동 및 화력 무기체계인 자주포, 전차, 장갑차를 기준으로 제대규모를 판단하였으며, 제대의 규모를 판단할 수 있는 특정 무기체계에 대한 탐지와 추적, 계수에 집중하였다. 모델은 객체탐지 알고리즘인 YOLOv8s를 객체 검출기, 다중 객체추적 알고리즘에서 높은 성능을 보여주는 BYTE 알고리즘을 추적기로 구성하여 객체탐지와 추적을 수행한다. 마지막으로 객체탐지와 추적에 이어 클래스별로 객체를 계수하고 계수된 값에 의해 규모를 판단하는 알고리즘을 추가하여 모델을 완성하였다. 학습 데이터는 지상에서 촬영된 이미지를 클래스별로 수집하였고 클래스는 자주포, 전차, 장갑차 3가지로 구성하였다. 모델 평가를 위해 제대규모를 반영한 테스트 영상으로 모델의 정확도를 확인하였으며 80%의 정확도를 나타내었다. 본 연구를 통해 향후 지상 감시장비를 활용해 군 제대규모를 판단할 수 있는 방법을 제시하였다는 것에 의의가 있다. The development of the Fourth Industrial Revolution, along with factors, such as population decline, will lead to a shift towards uncrewed systems in the future battlefield. The military will increasingly rely on uncrewed systems to carry out operations. Among them, uncrewed systems are replacing the surveillance and reconnaissance field. In this study, deep learning-based algorithms were used to judge the scale of ground-based military echelons moving in the surveillance area. A specific military unit was identified in the video using object detection and tracking techniques, and the scale of the military echelons was determined based on the count of detected objects. In particular, this study focused on detecting, tracking, and counting specific weapon units, such as self-propelled artillery, tanks, and armored vehicles. The model consists of the YOLOv8s object detection algorithm as an object detector and the BYTE algorithm as a multi-object tracking algorithm, showing high performance in object detection and tracking. Finally, the model was completed by adding algorithms that count objects by class and judge the size based on the counting. The training data was collected by the three types of classes from images taken from a ground perspective: self-propelled artillery, tanks, and armored vehicles. This study evaluated the model accuracy using test videos, and the result had an accuracy of 80%.

항공사진을 이용한 딥러닝 기반 FASCAM 투발 적합지역 분석

김기선(Kiseon Kim),마정목(Jungmok Ma) (사)한국CDE학회 2022 한국CDE학회 논문집 Vol.27 No.2

This study attempts to find FASCAM(Family of Scatterable Mines) areas automatically from aerial photographs. Currently, manual FASCAM area analysis is conducted in military, and it takes a long time and can generate inconsistent results. In order to select FASCAM areas automatically, an object classification of aerial photographs was first performed using deep learning techniques. DeepLabV3+ was selected as the deep learning model that showed good performance in the aerial photography classification. Areas with high similarity compared to sample images of FASCAM areas was selected as FASCAM areas. As the image similarity comparison method, the color histogram was used. Finally, the proposed FASCAM area analysis method showed 85.71% of agreement with military engineering officers. If the method of this study is used, it is expected to be helpful in analyzing the battlefield environment and establishing an obstacle plan.