| Photo Gallery | Original Image Download | |
|---|---|---|
| 01_Opening Ceremony | view | 01_Opening Ceremony.zip |
| 02_Plenary Talks | view | 02_Plenary Talks.zip |
| 03_Keynote Speeches | view | 03_Keynote Speeches.zip |
| 04_Young Researcher Sessions.zip | view | 04_Young Researcher Sessions.zip |
| 05_Best Paper Award Competition Session.zip | view | 05_Best Paper Award Competition Session.zip |
| 06_Exhibition | view | 06_Exhibition.zip |
| 07_Banquet | view | 07_Banquet.zip |
|
Korea Robotics Society (KROS) Privacy Policy Business Registration Certificate : 214-82-07990 / President : Hye-Kyung Cho Address : #506, The Korea Science and Technology Center, (635-4,Yeoksam-dong) 22, 7Gil, Teheran-ro, Gangnam-gu, Seoul, Korea SECRETARIATTEL +82-2-783-0306, FAX +82-2-783-0307, E-MAIL kros@kros.org Copyright © "Ubiquitous Robots 2021. |
| Hee-Sup Shin Northwestern University Bio-inspired soft strain sensing systems for measurement of wing deformation in small unmanned aerial vehicle |
|
Abstract Biological organisms demonstrate remarkable agility in complex environments, especially in comparison to engineered robotic systems. In part, this is due to an organism’s ability to detect disturbances and react to them quickly. Contrarily, small unmanned aerial vehicles (UAVs) often lose their flight stability in gusty environments. In this session, large-area soft strain sensing systems will be presented, which are designed to tackle the challenge of quickly sensing these disturbances on small UAVs in flight. Biography Hee-Sup Shin received the B.S. degree in mechanical engineering from Korea University, Seoul, South Korea, in 2013, and the M.S. and Ph.D. degrees in mechanical engineering from the Carnegie Mellon University, Pittsburgh, PA, USA, in 2015 and 2021, respectively. He is currently a postdoctoral research associate with the Querrey Simpson Institute for Bioelectronics, Northwestern University, Evanston, IL, USA. His research interests lie in soft sensors and actuators and their applications. |
| Christian Ott Technische Universität Wien, Austria Toward compliant robots that utilize their intrinsic body dynamics |
|
Abstract Inspired by the biological example of the human musculoskeletal system a large family of elastic actuator concepts have been proposed in the robotics research community. Still, robots driven by elastic actuators have not arrived at a maturity level as compared to state-of-the-art torque-controlled actuators, which nowadays are utilized in various commercial collaborative robots. While the general control properties of elastic robots are well understood, there is a gap in the realization of dynamic whole-body motions which can utilize the full capabilities of the elastic body dynamics. In the robotics literature whole-body motion generation and control has been dominated by the mastering of multi-body dynamics and either projection- or optimization-based methods. As a consequence, the resulting control actions can not be applied easily to the elastic robot dynamics and do not take the intrinsic elastic robot dynamics into account. In this talk I will give an overview of recent motion generation and control methods that aim at utilization of the natural dynamics and thus promise to allow a smoother application on elastic robots. A key concept in these methods will be the use of reduces models, which represent a subset of the robot’s motion capabilities exactly. This is in contrast to the concept of template models that represent an idealized behavior, which can only be approximated by the real robot. Such reduced models already have been successfully applied to dynamic locomotion tasks. The discussed ideas will be exemplified by experimental results with several torque controlled and elastic humanoid robots. Biography |
| Hyung-Soon Park KAIST, Korea What would be the best use of robots for neuro-rehabilitation? |
|
Abstract Since the late 1990s, rehabilitation robot has been a great model of medical robotics because the intrinsic characteristics of rehabilitation tasks matched well with the robotics technology available at the moment. Various exoskeleton and end-effector type rehabilitation robots have been developed for therapeutic and/or assistive tasks in upper limb and/or lower limb rehabilitation. There is no doubt that robotic devices are effective in orthopedic rehabilitation since it reduces human resources and enables more intensive and longer exercises. Regarding the neuro-rehabilitation which aims at recovery of musculoskeletal control after the injury in the nervous system, recent studies tell us rehabilitation robots are not necessarily superior to the conventional physical therapy. The key mechanism of neuro-rehabilitation might be represented by neuro-plasticity which has been the main interest in neurology. This presentation will discuss what would be the best use of robotics technology for promoting neuro-plasticity and introduce the recent research progress in the KAIST RENEW (Rehabilitation Engineering for Neurological disorders Worldwide) project. He was a research scientist at Rehabilitation Institute of Chicago from 2004 to 2009. From 2009 to 2013, he was a staff scientist with Rehabilitation Medicine Department at National Institutes of Health, Bethesda, MD. He joined KAIST in 2013 and he is now a professor in the Mechanical Engineering Department, Korea Advanced Institute of Science and Technology, Daejeon, Korea. He is directing research centers for future healthcare systems including KAIST global singularity project: RENEW (Rehabilitation Engineering for Neurological disorders Worldwide) and KAIST-CERAGEM center for future healthcare technology. His current research interest focuses mainly on optimizing robotics application for promoting brain plasticity in neuro-rehabilitation. |
| Kensuke Harada Osaka University, Japan Robotic Manipulation Research Aiming for Industrial Applications |
|
Abstract In this talk, we present our recent progress on robotic manipulation research aiming for industrial applications. We first explain why the automation of the high-mix and low-volume production is difficult. Then, we introduce several aspects of research mainly done by our research group, including task/motion planning, motion control, machine learning and gripper design. Biography |
| Jinah Jang POSTECH, Korea Bioprinted Human Tissues for Advanced Therapeutics |
|
Abstract Recent advances in biofabrication techniques have allowed for the fabrication of cardiac tissue models that are similar to the human heart in terms of their structure (e.g., volumetric scale and anatomy) and function (e.g., contractile and electrical properties). The importance of developing techniques for assessing the characteristics of 3D cardiac substitutes in real time without damaging their structures has also been emphasized. In particular, the heart has two primary mechanisms for transporting blood through the body: contractility and an electrical system based on intra- and extracellular calcium ion exchange. This talk will discuss how 3D cardiovascular tissue testing platform could be generated by integrating the concept of bioprinting-assisted tissue engineering and electrical sensing platforms. Combined with recent advances in human pluripotent stem cell technologies, printed human tissues could serve as an enabling platform for studying complex physiology in tissue and organ contexts of individuals. Biography |
| Sang-Youn Kim KoreaTech, Korea Soft Haptic Actuators and Sensors for Human Robot Interaction |
|
Abstract The term ‘haptic’ is a word that is related to kinesthetic or tactile sensation. Kinesthetic and tactile information refers to sensory data obtained through receptors of joints, muscles, ligaments, and etc, and through receptors of skin, respectively. A user recognizes the stiffness of an object through the kinesthetic information and discerns the texture of an object through the tactile information. Therefore, a user can communicate and/or interact with a robot efficiently by adding haptic information to auditory and visual information. Robots and their interfaces are under rapid shift from rigid to flexible and soft modules. Researchers are developing even shape changing interfaces, which can provide better affordance to users. Since such interface and module can have diverse shapes, the currently available rigid actuators/sensors is not very appropriate to provide/sense tactile feedback, and we will need the tactile actuators/sensors that have excellent shape conformity. Soft actuators/sensors are one of the best candidates for that purpose. This talk addresses the best-established technologies for soft haptic actuators and sensors. Biography |
| Min-hwan Oh Seoul National University Randomized Exploration in Structured Reinforcement Learning |
|
Abstract Recent years have witnessed increasing empirical successes in reinforcement learning (RL). Yet, we still have fundamental questions that are not well understood in RL. For example, how many observations are required for a decision-making RL agent to learn how to act optimally? How can the agent explore efficiently in feature space? Common approaches to exploration are highly inefficient. I will discuss how this can be addressed with a structured Markov decision process and randomized exploration, which enables efficient exploration and learning in feature space and allows RL to be data-efficient and practical. Biography |
| Guy Williams University of Tasmania, Australia The new golden era of polar research and exploration with autonomous systems |
|
Abstract As climate change becomes humanity’s greatest existential crisis, the race is on to convince the world to take urgent action. A key factor in motivating this action is a clear understanding of how the global climate system works and specifically our ability to monitor its current state and model its future. Polar regions are at the same time critical components of this system and some of the most poorly observed, due to the extreme polar conditions and challenging logistics. This has resulted in persistent data and knowledge gaps – gaps that have existed for decades and continue to negatively impact climate action through the uncertainty they inadvertently feed. Robots, or autonomous systems, present more than just an exciting and innovative ‘tool’ for Antarctic science – they are quite possibly the only chance we have to fill these critical gaps and improve the accuracy of vital metrics such as the impact of polar melting on global sea-level rise. But simply adding robots to polar research expeditions doesn’t make all your problems go away – if anything, your challenges will increase. Nonetheless, a new golden era of polar exploration awaits those groups with the courage and expertise to utilise autonomous systems in this high risk/high reward research field. In this talk I’ll discuss the great progress made this century with polar autonomous systems, some of the lessons learnt and how we need to work together as scientists and engineers, within and across national borders, to observe, monitor and protect our precious polar regions. Biography |
| Toshio FUKUDA Nagoya University and Waseda University, Japan AI Robots and Moon Shot Program |
|
Abstract There are many ways to make research and development of intelligent robotic systems. I have been working on the Multi-scale robotics systems for many years, based on the Cellular Robotics System, which is the basic concept of the emergence of intelligence in the multi-scale way from Organizational Level, Distributed robotics to Biological Cell engineering and Nano-robotics with self-reconfigurability. It consists of many elements how the system can be structured from the individual to the group/society levels in analogy with the biological system. Focusing on the coevolution and self organization capabilities, I will show a new initiative on AI and Robot, one of the Moon Shot Programs started by Japanese Government, since 2020. Based on the Society 5.0, it is a new and challenging program aiming at the AI robotic system in 2050. I will introduce some of the projects in this program for realization of the Society 5.0 by back-casting technologies from the 2050 to the current ones. Biography |
| Jiyoun Moon Chosun University Semantic scene understanding based human-robot cooperation |
|
Abstract Human-robot cooperation is unavoidable in various applications ranging from manufacturing to field robotics owing to the advantages of adaptability and high flexibility. Especially, complex task planning in large, unconstructed, and uncertain environments can employ the complementary capabilities of human and diverse robots. For a team to be effectives, knowledge regarding team goals and current situation needs to be effectively shared as they affect decision making. In this respect, semantic scene understanding in natural language is one of the most fundamental components for information sharing between humans and heterogeneous robots, as robots can perceive the surrounding environment in a form that both humans and other robots can understand. In this presentation, semantic scene understanding based human-robot cooperation is introduced. Biography Jiyoun Moon received the B.S degree in Robotics from Kwangwoon University in 2014, and the Ph.D. degree in Electrical and Computer Engineering from Seoul National University in 2020. She is currently an assistant professor at Chosun University. Her research interests include artificial general intelligence, cognitive robotics, and human-robot cooperation. |
| Seokju Lee Korea Institute of Energy Technology (KENTECH), Korea Computer Vision Meets Energy AI: From Autonomous Driving to Carbon Neutrality |
|
Abstract Computer vision is indispensable for existing/emerging AI industries, and its development is accelerated by the advances in machine learning. In particular, autonomous driving technology is the culmination of these studies. Recently, automobiles are facing another turning point in this era of electric vehicles. In this presentation, I will present the future of automobiles as autonomous energy mobility and my vision at KENTECH. Contents: |
| Jeong-Jung Kim Korea Institute of Machinery & Materials, Korea AI for Mobile Manipulation in Unstructured Environments |
|
Abstract A mobile manipulator is a robot system combined with a manipulator on a mobile robot. Since the system has a high degree of freedom and a wide workspace, it can be used for various task in typical human environments. Until now, the system is mainly used for simple tasks, and in order to be used in unstructured environment, it should be able to handle the manipulation of various objects and navigation in narrow passages. In this presentation, I will show how these problems are addressed with AI for the mobile manipulation. Biography Jeong-Jung Kim received the B.S degree in Electronics and Information Engineering from Chonbuk National University in 2006, and the M.S. degree in Robotics and the Ph.D. degree in Electrical Engineering from Korea Advanced Institute of Science and Technology (KAIST) in 2008 and 2015, respectively. From 2015 to 2017, he was a post-doctoral researcher at Robotics and Media Institute in Korea Institute of Science and Technology (KIST). He is currently a senior researcher at Korea Institute of Machinery & Materials (KIMM). His research interests include AI for manipulation and navigation in robotics. |
| Myunghee Kim University of Illinois at Chicago, USA Human-Wearable robot Co-adaptation |
|
Abstract Reduced mobility is a significant societal problem. In 2010, 30.6 million adults had ambulatory limitations, and 23.9 million individuals found it difficult to walk one-quarter mile, only in US alone. This reduced mobility is related to increased medical problems and can have detrimental socio-economical impacts. The problem will only increase with an aging population. The drive to discover effective strategies for human gait assistance has led to investigations into human-wearable robots. In response, my research strives to advance the field by focusing on wearable robots that can respond to individual users, resulting in a smart assistance strategy in which the robot adapts to the human wearers. Furthermore, my research includes the concept of co-adaptation, in which the human users receive guidance on how to adjust their movement patterns to optimize their benefit from the personalized robot. In this talk, I will introduce a robot adaption method to a user, human-in-the-loop (HIL) optimization, and the user guidance method to facilitate robot use. The HIL optimization is a machine learning approach using biofeedback, which significantly reduced walking and squatting efforts when users wore various wearable-robots such as hip and ankle soft exosuit, ankle-foot orthosis, and ankle-foot prosthesis. My group also found that user guidance via visual feedback can improve the wearable robot use, even for an unfavorable condition – initially increased the cost of walking. The seminar will be concluded with a discussion about the challenges and opportunities offered by the human-in-the-loop assistance controller and user guidance method. Biography |
| Inwon Jong Fren Inc. Chief Design Officer, USA We design healing environments by bringing purposeful stories to life. |
|
Abstract We identify ways to personalize experiences and optimize services by transforming our understanding into creative storytelling environments. We also utilize co-creation to develop guided storytelling through user experience and designed touchpoints. The outcome presents a healing environment that incorporates educational and therapeutic activities designed to engage and support inpatient care with extraordinary experiences. Our process behind a project combines service and business design capabilities. In addition to the usual customer research, we make an effort at an early stage of the project, to understand the business as well as the organization to create boundaries for the project. The deliverables are designed to serve the customers and users, the business, and the organization. Biography Inwon Jong is a Savannah, Georgia-based designer having affinities with design and technology. His work relies on the bridge between the areas of digital aesthetics and user experiences which explore the expressive use of a computational method. Inwon has served as an accomplished product designer in the field of Information Technology since 2005. Over 10 years of his design career was spent with Samsung Electronics, which had a clear and direct impact on the interaction software utilized by mobile phones and worked with a diverse range of co-workers across multiple mobile platforms. He is currently a Chief Design Officer at Frendesign where he focuses on designing and developing programs that yield digital artifacts and environments for therapeutic installations that directly improve children’s patient experiences. |
| Deokjin Lee Jeonbuk National University, Korea Model-free vs Model-based Reinforcement Learning to Control for Robotics |
|
Abstract After AlphaGo vs Lee Sedol in 2016, known as a surprise 4-1 victory of the Google DeepMind Challenge Match in Seoul, South Korea, the topic of reinforcement learning (RL) has received huge attention for this breakthrough while leading to a new level of artificial intelligence. In general, reinforcement learning methods fall into two categories, model-based method and model-free method, and each of which shows unique advantages. Even though the rich theoretical foundation of model-free deep reinforcement learning (DRL) and their various application tasks, RL requires many data samples to find optimal policies to realize good performance, and explore all possible actions which may be subject to the unstable safety issue. While, model-based RL can quickly obtain near-optimal control by learning the model in a rather limited class of dynamics. However, its disadvantages lie on the fact that most model-based algorithms learn local models over-fitting several samples by depending on simple functional approximators. In this talk, we are going to compare the RL algorithms using in both model-based and model-free approaches with a realistic robot control applications, such as drone stabilization and quadruped robot control, investigating how RL agents generalize to real-world autonomous robot control-tasks. In addition, a new hybrid approach which integrates the model-based and model-free RLs is discussed with a robot control application. Biography received the Ph.D. & MS degrees in Aerospace Engineering from Texas A&M University in May 2005. BS in Mechanical Aerospace Systems from Jeonbuk National University in 1996. He worked for Agency for Defense Development (ADD) from 2006 to 2007, and from 2009 to 2011 he was also a research professor at the Center for Autonomous Vehicle Research (CAVR), Naval Postgraduate School, Monterey, CA, U.S.A. Currently, he is an associate professor at the School of Mechanical Design Engineering at Jeonbuk National University, also a director of the center of Autonomous Intelligence e-Mobility(CAIM). IEEE Senior Member (2015~), AIAA Senior Member(2010~), Texas A&M University Excellent Research Fellowship (2005), The John V. Breakwell Student Travel Award(AAS, 2003), President of KSME Artificial Intelligence Research Society(2020~2021), Institute of Control, Robotics & Systems(ICROS), Best Paper AIAA(2008), Best Paper KSME (2016, 2017), etc. |
| Min Jun Kim Southern Methodist University, USA Magnetically Actuated Modular Robots for Self-Assembling and Additive Manufacturing |
|
Abstract Magnetically actuated modular robots can be controlled remotely by external magnetic fields, making them promising candidates for biomedical and engineering applications. This talk will introduce an innovative reconfigurable modular robotic system which controls miniature components that can be actively assembled and disassembled on command. This type of system could potentially improve the robustness and control-lability of small-scale additive manufacturing. The base components are miniature cu-bes that contain permanent magnets. They are actuated using an external magnetic field generated via a three axis Helmholtz coil system. The cubes can achieve different motion patterns such as such as pivot walking, tapping, and tumbling. Our project in-volves designing and fabricating scalable modular subunits using 3D printing. A set of design rules for the cubes has been defined. Algorithms to control the magnetic subu-nits have been studied. The issues addressed by this talk are at the interface of small-scale robotics, control theory, materials science, and bioengineering, and hold exciting prospects for fundamental research with the potential for diverse applications. Biography Dr. MinJun Kim is presently the Robert C. Womack Endowed Chair Professor of En-gineering at the Department of Mechanical Engineering, Southern Methodist Universi-ty. He received his B.S. and M.S. degrees in Mechanical Engineering from Yonsei University in Korea and Texas A&M University, respectively. Dr. Kim completed his Ph.D. degree in Engineering at Brown University, where he held the prestigious Si-mon Ostrach Fellowship. Following his graduate studies, Dr. Kim was a postdoctoral research fellow at the Rowland Institute in Harvard University. He joined Drexel University in 2006 as Assistant Professor and was later promoted to Professor of Me-chanical Engineering and Mechanics. Dr. Kim has been exploring biologically inspired sensing and actuation to develop new types of nano/micro robotics. His notable awards include the National Science Foundation CAREER Award (2008), Drexel Ca-reer Development Award (2008), Human Frontier Science Program Young Investiga-tor Award (2009), Army Research Office Young Investigator Award (2010), Alexan-der von Humboldt Fellowship (2011), KOFST Brain Pool Fellowship (2013 & 2015), Bionic Engineering Outstanding Contribution Award (2013), Louis & Bessie Stein Fellowship (2008 & 2014), ISBE Fellow (2014), ASME Fellow (2014), Top10 UNESCO-Netexplo Award (2016), KSEA & KOFST Engineer of the Year Award (2016), IEEE Senior Member (2017), Sam Taylor Fellowship (2018), Gerald J. Ford Research Fellowship (2018), and Protégé of the Academy of Medicine, Engineering and Science of Texas (2019). |