Ung Hee Lee - Research

Improving human mobility using robotics, and machine learning

Wearable robotics, including powered exoskeletons and prosthetics, hold immense potential to enhance quality of life by reducing physical strain and improving mobility. However, despite the rapid advancements in wearable devices and microcontrollers, their adoption in daily life remains limited. A significant barrier lies in developing control systems capable of bridging the gap between research and real-world application.

To address this challenge, my research focuses on creating a generalizable control policy that adapts to various tasks and individual wearers. Central to this vision is the development of a multimodal machine learning framework for human-robot interaction, enabling effective transferability across different users and environments. My program is anchored by two primary objectives: (1) building a foundational control model for wearable robots, and (2) optimizing task-specific and user-specific control policies.

I believe these foundational components are critical for wearable robotics to provide seamless assistance across diverse environments while accommodating individual user preferences. By leveraging my expertise in mechatronics, biomechanics, and machine learning, my research aims to establish a robust, data-driven foundation for wearable robotics. This approach has the potential to significantly advance their widespread integration into real-world applications, making this transformative technology accessible and practical.

Intelligent and Autonomous Wearable Robotic System

Recent research in wearable robotics has made significant strides in personalizing and optimizing lower-limb assistive devices through innovative data-driven approaches and advanced machine learning techniques. Our work explores the intersection of human biomechanics, machine learning, and robotic assistance, focusing on developing intelligent systems that adapt to individual user preferences and locomotor intentions. By employing sample-efficient active learning methods, we have created novel strategies for optimizing ankle exoskeleton control, enabling more intuitive and personalized device performance. Additionally, we have predicted preferred ankle stiffness in quasi-passive prostheses and utilized image transformation techniques with convolutional neural networks to encode human locomotor intent. This comprehensive approach enhances wearable robotic technologies, demonstrating how advanced computational methods can bridge the gap between human physiological needs and robotic assistance. Ultimately, our research aims to create more autonomous, responsive, and user-centric assistive devices that dynamically adjust to individual biomechanical characteristics and movement patterns, thereby improving rehabilitation outcomes, mobility assistance, and overall quality of life for individuals requiring lower-limb prosthetic or orthotic support.

Relevant publications

U Lee, V Shetty, P Franks, J Tan, G Evangelopoulos, S Ha, E J Rouse. User Preference Optimization for Control of Ankle Exoskeletons using Sample Efficient Active Learning. Science Robotics 2023.

V Shetty, U Lee, K Ingraham, E J Rouse. A Data Driven Approach for Predicting Preferred Ankle Stiffness of a Quasi-Passive Prosthesis. IEEE Robotics and Automation Letters with Presentation at International Conference on Robotics and Automation 2022.
U Lee, J Bi, R Patel, D Fouhey, and E J Rouse. Image Transformation and CNNs: A Strategy for Encoding Human Locomotor Intent for Autonomous Wearable Robots. IEEE Robotics and Automation Letters with Presentation at International Conference on Intelligent Robots and Systems 2020.

Electric Motors for Lightweight Robotics

Recently, brushless motors with especially high torque densities have been developed for applications in autonomous aerial vehicles (i.e. drones). These motors are promising for other applications, such as humanoids and wearable robots; however, the emerging companies that produce motors for drone applications do not typically provide adequate technical specifications that would permit their general use across robotics. My research encompasses characterizing these motors, and guide selection process for optimal design of lightweight robots.

Relevant publications

U Lee, C Pan, and E J Rouse. Empirical Characterization of a High-performance Exterior-rotor Type Brushless DC Motor and Drive. IEEE International Conference on Intelligent Robots and Systems 2019.
U Lee, T. Shepherd, S. Kim, A. De, H. Su, R. Gregg, E. Rouse. How to Model Brushless Electric Motors for the Design of Lightweight Robotic Systems. Arxiv submission. Under review.