Daehyung Park is a Ph.D. candidate in the College of Computing and the Institute for Robotics and Intelligent Machines (IRIM) at the Georgia Institute of Technology, working with Dr. Charles C. Kemp in the Healthcare Robotics Laboratory. He is currently researching assistive manipulation and multimodal execution monitoring methods for people with disabilities. 

Park has experience in a broad range of robotics in both academic and industrial areas. He received a B.S. at Osaka University in Japan, where he researched HRP-2' turning motion (advisor: Dr. Arai Tatsuo). He received an M.S. at the University of Southern California, where he researched movement primitives for articulated manipulators (advisor: Dr. Stefan Schaal). After earning his master’s degree, he worked as a robotics researcher at the Mechatronics R&D Center of Samsung Electronics Inc.



Research Interests  -  Manipulation, Machine Learning, Perception


News
  • My IROS video in IEEE spectrum's Video Friday
  • D. Park, Y. Hoshi, and C. C. Kemp. “A Robot-Assisted Feeding System using a General Purpose Mobile Manipulator”, (Under preparation)
  • D. Park, Y. Hoshi, and C. C. Kemp. “A Multimodal Anomaly Detector for Robot-Assisted Feeding Using an LSTM-based Variational Autoencoder”, IEEE Robotics and Automation Letters (RA-L) (Under review)
  • D. Park, H. Kim, and C. C. Kemp. “Multimodal Anomaly Detection for Assistive Robots”, Autonomous Robots (Under review)
  • D. Park, H. Kim, Y. Hoshi, Z. Erickson, A. Kapusta, and C. C. Kemp. “A Multimodal Execution Monitor with Anomaly Classification for Robot-Assisted Feeding”, IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS2017) 
Recent Projects
A Multimodal Execution Monitor

We introduce a new execution monitor that detects and classifies anomalies using multimodal sensory signals for manipulation tasks. The monitor models the spatiotemporal dynamics of the sensor information using a generative model (i.e., hidden Markov models or long short-term memory networks). The monitor detects an anomaly when current anomaly score is higher than a state-based threshold. The monitor then classifies the type and cause of an anomaly using mutilayer perceptron (MLP).



feeding_feature

Robot-Assisted Feeding System

We present a proof-of-concept robotic system for assistive feeding using a general-purpose mobile manipulator for people with disabilities.



haptic_perception

Haptically-guided Mapping, Planning, and Control

In this work, we are focusing on methods for haptic mapping, planning, and control during reaching into the unknown environment.