[SME – NCKH] SEMINAR NGHIÊN CỨU KHOA HỌC THÁNG 3

1.  Title: “AI Research Activities in Robotics RL Lab” – Prof. Jeha Ryu 

Google Scholar Profile: https://scholar.google.com/citations?user=zHoKd10AAAAJ&hl=ko

        AI (Artificial Intelligence) is everywhere these days and will prevail everywhere, every time, everybody in all areas of our lives. Especially, we can see and experience AI in all machines and artificial creatures like robotic systems and metaverse. This means that we all engineers including mechanical and mechatronic engineers must study and research AI. This small talk will briefly introduce the AI research activities in Robotics AI Lab, IIT, at GIST: 1) Deep-Learning-Based Emergency Stop Prediction for Robotic Lower-Limb Rehabilitation Training Systems, 2) DeepLearning-Based Registration of Diagnostic Angiogram and Live Fluoroscopy for Percutaneous Coronary Intervention, 3) Learning an Accurate State Transition Dynamics Model by Fitting Both a Function and its Derivative, 4) Deep Learning-Based Subtask Segmentation of Timed Up-and-Go Test Using RGB-D Cameras, 5) A New Approach to Traffic Accident Anticipation With Geometric Features for Better Generalizability, 6) Voronoi Tessellation for Efficient Sampling in Gaussian Process-Based Robotic Motion Planning, etc.

2. Title: “Rehabilitation robotics and nanorobotics for targeted therapy of brain disorders” – Prof. Jungwon Yoon 

Google Scholar Profile:https://scholar.google.com/citations?user=oG-utS8AAAAJ&hl=en

        Abstract:
For rehabilitation robotics, one of the primary goals of stroke rehabilitation is the treatment of gait abnormalities. Numerous studies and associated technologies, such as wearable and treadmill-based exoskeleton robots, are being proposed. Still, there are high demands that outperform conventional rehabilitation techniques in terms of effectiveness and cost. In this invited talk, we propose a haptic biofeedback training scheme to improve balance, gait speed, gait symmetry, and muscle activation for trained subjects, either using haptic mode alone or in conjunction with robot devices. By using a fingertip contact with a force of less than one Newton, the Light Touch (LT) allows healthy users to obtain reference frame data that indicates the location and orientation of their present body centre. It supports balance instinctively and has been shown to be useful in the field of neuroscience both statically and dynamically. Thus, basic haptic biofeedback using LT can enhance patients’ sense of balance, make up for reduced proprioception brought on by brain disorders like stroke, and allow patients to control their walking speed on level ground without the need for bulky exoskeletons. It has been confirmed that haptic bio-feedback systems, including kinesthetic and tactile wearable haptic devices, can improve stroke survivors’ balance, gait symmetry, gait speed, and muscle activation at the paralyzed limb. Furthermore, haptic integrated training scheme using a rehabilitation robot was investigated. When compared to current rehabilitation robot devices, the haptic training scheme is more affordable and facilitates rehabilitation training more effectively. It also enables the patient to wear and effectively execute gait and balance training. It is anticipated to be a novel perspective on rehabilitation therapy that can overcome its current drawbacks and work in combination with other rehabilitation robots now in use. More work needs to be done on developing different haptic integrated training modes, like upper limb reaching training and upper and lower limb coordination training, which are also easily available in wearable form. For nanorobotics, magnetic nanoparticles (MNPs) are a promising candidate for use as carriers in targeted drug delivery systems because they can function at both the cellular and molecular levels. Electromagnetic sensing and guidance schemes using magnetic nanoparticles (MNPs) can allow a nanotechnology-based drug delivery approach to be feasible for targeted therapies for brain diseases such as brain cancer, stroke, and Alzheimer’s disease. Magnetic particle imaging (MPI) is a fast and sensitive imaging modality that is used to measure the spatial distribution of MNPs. MPI systems offer spatial resolutions on the millimeter scale and high temporal resolutions, which fulfill the requirements for cardiovascular, neurological, and peripheral vascular applications. An electromagnetic navigation scheme using MPI can deliver magnetic nanoparticles to efficiently targeted regions of a brain with feedback information while minimizing particles’ aggregation and passing through blood brain barrier (BBB). This talk will show how MPI scheme can be combined together with the electromagnetic guidance scheme. The proposed MPI-based targeting approaches can be finally adapted to medical robotic platforms for brain drug targeting, brain stimulation, and brain hyperthermia. We expect the proposed combined robotic scheme with target drug delivery and target rehabilitation training will provide a new therapy option in stroke disease.