Haodong Chen

Haodong Chen

Postdoctoral Associate

Robotic Cognitive Rehabilitation System for Mild Cognitive Impairment

Robotic Rehabilitation System

Goals

  • Develop a Robotic Cognitive Rehabilitation System to enhance cognitive training for patients with Mild Cognitive Impairment (MCI).
  • Implement Advanced Target Acquisition Algorithms using computer vision to accurately detect and guide cognitive tasks.
  • Create an Immersive One-to-Many Rehabilitation Model that allows a single physician to monitor and assist multiple patients simultaneously through intuitive graphical user interfaces (GUIs).
Rehabilitation Goals

Key Findings

  • Target Acquisition Algorithms: Developed modulus-shift matching (MSM) and Fourier descriptor (FD) algorithms to achieve precise detection and positioning of cognitive task targets with less than 2° error.
  • Graphical User Interfaces (GUIs): Designed two distinct GUIs for physicians (I-GUI) and patients (P-GUI), enabling efficient task assignment, real-time monitoring, and intuitive patient interaction.
  • One-to-Many Rehabilitation Model: Established a scalable rehabilitation framework where a single physician can oversee multiple patients concurrently, significantly improving resource utilization.
  • Robotic Arm Integration: Integrated a six degrees-of-freedom robotic arm controlled via YJCTRL-A601 to provide real-time demonstrations and assistance during cognitive tasks.
  • Automated Data Management: Implemented end-to-end software solutions in MATLAB and C++ for automated task setup, progress tracking, and data storage, enhancing the efficiency of cognitive rehabilitation sessions.
Key Contributions

Technologies Utilized

  • Computer Vision: Utilized Canny edge detection, Gaussian convolutions, and Fourier descriptors for accurate target acquisition and position extraction.
  • Machine Learning: Employed unsupervised learning techniques such as spectral clustering and Gaussian Mixture Models (GMM) to optimize motion patterns and improve system efficiency.
  • Software Development: Developed robust software using MATLAB, C++, and Python to handle image processing, algorithm implementation, and GUI functionalities.
  • Robotics: Integrated Delta Robots and the Robotic Operating System (ROS) for real-time control and seamless human-robot collaboration.
  • User Interface Design: Created intuitive GUIs for both physicians and patients to facilitate easy task management and real-time feedback.

Impact

This Robotic Cognitive Rehabilitation System significantly enhances the efficiency and effectiveness of cognitive rehabilitation for patients with MCI. By automating target acquisition and providing real-time robotic assistance, the system reduces the reliance on manual intervention, allowing physicians to manage multiple patients simultaneously. The integration of computer vision and machine learning ensures high accuracy in task execution, thereby improving patient outcomes. Additionally, the immersive and interactive nature of the system fosters better patient engagement and adherence to rehabilitation protocols, ultimately contributing to advancements in human-robot collaboration and cognitive therapy technologies.

Selected Publications

  1. Chen, H.-D., Wang, Y.-F., Guo, Z., Chen, W.-X., & Zhao, P. (2018). A GUI Software for Automatic Assembly Based on Machine Vision. IEEE International Conference on Mechatronics, Robotics and Automation (ICMRA), Hefei, China, May 18–21, 105–111.
  2. Chen, H., Teng, Z., Guo, Z., & Zhao, P. (2020). An Integrated Target Acquisition Approach and Graphical User Interface Tool for Parallel Manipulator Assembly. ASME J. Comput. Inf. Sci. Eng., 20(2), 021006. parative Study of Curvature Scale Space and Fourier Descriptors for Shape-Based Image Retrieval. J. Visual Commun. Image Representation, 14(1), 39–57.
  3. Chen, H.D., Zhu, H., Teng, Z. and Zhao, P., 2020. Design of a robotic rehabilitation system for mild cognitive impairment based on computer vision. Journal of Engineering and Science in Medical Diagnostics and Therapy, 3(2), p.021108.