• The Journal of Korea Robotics Society
• /
• v.17 no.1
• /
• pp.76-85
• /
• 2022

One of the most fundamental challenges when designing controllers for dynamic systems is the adjustment of controller parameters. Usually the system model is used to get the initial controller, but eventually the controller parameters must be manually adjusted in the real system to achieve the best performance. To avoid this manual tuning step, data-driven methods such as machine learning were used. Recently, reinforcement learning became one alternative of this problem to be considered as an agent learns policies in large state space with trial-and-error Markov Decision Process (MDP) which is widely used in the field of robotics. However, on initial training step, as an agent tries to explore to the new state space with random action and acts directly on the controller parameters in real systems, MDP can lead the system safety-critical system failures. Therefore, the issue of 'safe exploration' became important. In this paper we meet 'safe exploration' condition with Control Barrier Function (CBF) which converts direct constraints on the state space to the implicit constraint of the control inputs. Given an initial low-performance controller, it automatically optimizes the parameters of the control law while ensuring safety by the CBF so that the agent can learn how to predict and control unknown and often stochastic environments. Simulation results on a quadrotor UAV indicate that the proposed method can safely optimize controller parameters quickly and automatically.

• The Journal of the Korea Contents Association
• /
• v.20 no.4
• /
• pp.305-314
• /
• 2020

Mobile environment is relatively free from time and space constraints, and has unique characteristics such as not applying physical and social vulnerability because of it's non-face-to-face channel. Since such spatial and temporal speciality is likely to appear in a different form from psychological action and action in a general face-to-face environment, it is necessary to conduct an empirical study based on the results of the existing research that has been studied as major psychological variables of action in mobile environment. However, advance research have not been conducted in earnest on the vulnerability and anxiety of the consumers in a new environment called mobile finance. so this study intends to establish the basis through empirical research. As a result, women among traditional vulnerable groups perceived themselves to be vulnerable in mobile financial environments than men and felt fear of crime strongly. In addition, the vulnerability of mobile financial crime did not have a significant direct effect on mobile financial use, but it was confirmed that mobile crime fear was mediated completely and had an indirect effect on mobile financial use.

• Journal of The Korean Association For Science Education
• /
• v.41 no.2
• /
• pp.115-131
• /
• 2021

As the global climate change emergency is escalating, the need for 'Social Action-Oriented SSI (SAO-SSI) on climate change topics' in science education that can change society through social activity is increasing. By employing sociocultural theory, this study explores the challenges of limiting teacher's agency in implementing SAO-SSI on climate change topics in science education. Data from participant observation for 46 lessons, in-depth interviews with participants, field notes, and teacher reflection notes were analyzed by the structure of into micro- (classrooms), meso- (school), and macro- (Korea society) level. At the micro-level, the teacher's new attempts of SAO-SSI on climate change topics class made it difficult for him to identify students' understanding of climate change, because they have a low sense of perception that climate change is also their problem. In addition, the teacher had difficulties leading students' into an engagement for social action because students were skeptical about the feasibility of planned social behavior by positioning themselves as children or had difficulty in understanding social action and sympathizing with its values. At the meso-level, a school culture that encourages the implementation of a curriculum similar to that of colleagues, it was difficult to implement one's own curriculum. And it was difficult to develop expertise without the support and communications with colleagues who revealed the burden of unfamiliar science topics of climate change. In addition, conflicts arose in the process of implementing out-of-school social actions with the principal's passive support. At the macro-level, the insufficient proper material resources for SAO-SSI on climate change topics class, and negative perceptions on the students' social action in the society were acting as constraints. We offer implications for what kind of structural support and efforts from various subjects in the educational community should be provided to implement SAO-SSI on climate change topics class in science education.

• Journal of KIISE:Software and Applications
• /
• v.35 no.12
• /
• pp.780-788
• /
• 2008

Modeling hand poses and tracking its movement are one of the challenging problems in computer vision. There are two typical approaches for the reconstruction of hand poses in 3D, depending on the number of cameras from which images are captured. One is to capture images from multiple cameras or a stereo camera. The other is to capture images from a single camera. The former approach is relatively limited, because of the environmental constraints for setting up multiple cameras. In this paper we propose a method of reconstructing 3D hand poses from a 2D input image sequence captured from a single camera by means of Belief Propagation in a graphical model and recognizing a finger clicking motion using a hidden Markov model. We define a graphical model with hidden nodes representing joints of a hand, and observable nodes with the features extracted from a 2D input image sequence. To track hand poses in 3D, we use a Belief Propagation algorithm, which provides a robust and unified framework for inference in a graphical model. From the estimated 3D hand pose we extract the information for each finger's motion, which is then fed into a hidden Markov model. To recognize natural finger actions, we consider the movements of all the fingers to recognize a single finger's action. We applied the proposed method to a virtual keypad system and the result showed a high recognition rate of 94.66% with 300 test data.

• Journal of the Korean Society of Marine Environment & Safety
• /
• v.25 no.2
• /
• pp.178-188
• /
• 2019

When using a distributed system, it is very important to know the intention of a target ship in order to prevent collisions. The action taken by a certain ship for collision avoidance and the action of the target ship it intends to avoid influence each other. However, it is difficult to establish a collision avoidance plan in consideration of multiple-ship situations for this reason. To solve this problem, a Distributed Stochastic Search Algorithm (DSSA) has been proposed. A DSSA searches for a course that can most reduce cost through repeated information exchange with target ships, and then indicates whether the current course should be maintained or a new course should be chosen according to probability and constraints. However, it has not been proven how the parameters used in DSSA affect collision avoidance actions. Therefore, in this paper, I have investigated the effect of the parameters and weight factors of DSSA. Experiments were conducted by combining parameters (time window, safe domain, detection range) and weight factors for encounters of two ships in head-on, crossing, and overtaking situations. A total of 24,000 experiments were conducted: 8,000 iterations for each situation. As a result, no collision occurred in any experiment conducted using DSSA. Costs have been shown to increase if a ship gives a large weight to its destination, i.e., takes selfish behavior. The more lasting the expected position of the target ship, the smaller the sailing distance and the number of message exchanges. The larger the detection range, the safer the interaction.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.33 no.4B
• /
• pp.207-214
• /
• 2008

Under future internet environment, wireless sensor networks will be used in a wide range of applications. A major problem for designing sensor protocol is developing the most energy efficient technique to monitor an area of interest for a long time since sensors have some constraints such as small and a limited energy level. In addition, data latency is often a critical issue since sensory data is transmitted via multi hop fashion and need to be delivered timely for taking an appropriate action. Our motivation for designing a data forwarding protocol is to minimize energy consumption while keeping data latency bound in wireless sensor networks. In this paper, we propose a data forwarding protocol that consists of wakeup scheduling and MAC protocols, the latter of which is designed to achieve load balancing. Simulation results show that the proposed framework provides more energy-efficient delivery than other protocol.

• Journal of Korean Society of Industrial and Systems Engineering
• /
• v.40 no.1
• /
• pp.11-21
• /
• 2017

Artillery fire power due to effectiveness which is hard to predict well-planned and surprising attack can give a fear and shock to the personnel and is a very core weapon system and takes a critical role in wartime. Therefore in order to maximize operational effectiveness, Army required protecting artillery and takes a quick attack action through rapid construction of artillery's positions. The artillery use artillery's position to prevent exposure by moving to other position frequently. They have to move and construct at new artillery's positions quickly against exposing existed place by foe's recognition. These positions should be built by not manpower but engineering construction equipment. Because artillery positions have to protect human and artillery equipment well and build quickly. Military engineering battalion have lots of construction equipment which include excavator, loader, dozer, combat multi-purposed excavator, armored combat earthmover dump truck and so on. So they have to decide to optimal number of Team combining these equipments and determine construction sequence of artillery's position in operational plan. In this research, we propose to decide number of Team efficiently and allocate required construction's positions for each Team under constraints of limited equipments and time. To do so, we develop efficient heuristic method which can give near optimal solution and be applied to various situation including commander's intention, artillery position's priority or grouping etc. This heuristic can support quick and flexible construction plan of artillery positions not only for using various composition's equipment to organize Teams but also for changing quantity of positions.

• Journal of KIISE:Software and Applications
• /
• v.34 no.8
• /
• pp.763-771
• /
• 2007

As software is widely used on various environments today, there is an increasing need for adaptive software. Adaptive software is robust and flexible software that modifies its own behavior in response to the changes in its environment. Due to time constraints, high complexity and so on, it is hard to acquire context information from environment. So, when implementing software, it is common to think easily acquired data to be the environments. This research proposes an adaptive software framework that assumes plane images to be environments. Plane images are easy to acquire and have enough complexity. From this, our framework is able to acquire context information, reasons with action rule, and acts on the result of reasoning. Stand on this framework, implements software that plays a simple game automatically.

• Journal of the Korean Society for Railway
• /
• v.15 no.5
• /
• pp.459-466
• /
• 2012

Due to the issues such as omission of data, document based management, maintenance based on measurement data and wire-based network, it is difficult existing maintenance system for railway bridges to act to diverse characteristics of site and environmental changes in real time. With these reasons, there are many constraints in establishing active maintenance strategies for railway bridges. To solve these issues, this study suggests an integrated maintenance business model based on practical utilization and information management based on BIM technology to build a smart maintenance operation system based on ubiquitous computing for railway bridges. To secure its development and practical applications, a quantitative evaluation by questionnaire analysis was performed. Therefore, it is expected that the suggested model will be utilized as a framework model in order to build the smart maintenance operation system from collection of maintenance data to action for railway bridges.

• The Journal of the Korea Contents Association
• /
• v.19 no.7
• /
• pp.110-119
• /
• 2019

is a movie about the frustration and the despair of young people in Korean society today. It is different from the previous Lee Chang-dong's film in that it provides the main character's perception and feelings about the world as the ultimate cause of the protagonist's actions. Lee Chang-dong's characters begin to change from . The main character of Burning goes one step further. He seems to believe that the main character's only weapon to overcome the logic and constraints of reality is obsessive and compulsive behaviors. Unlike previous movies, the character in Burning is a person who practices awareness and proves his character through action. I will analyze characters in to suggest the change of film form. Especially I will look at how the character's actions and the writer's own voice can meet and what results it bring about in this movie. It will lead to his changes in film form and an idea on the young people in Korean society today.