Ergonomically Intelligent Teleoperation and Physical Human-Robot Interaction

Details

• Personal Author:
  Yazdani M
• Description:
  Workplace injury is a concerning issue affecting millions of workers worldwide annually. Work-related musculoskeletal disorders (WMSDs) are the second most significant cause of disabilities worldwide, and awkward postures are known to be the main contributor WMSDs. Robotics, teleportation, and human-robot collaboration are well-suited alternatives for high-risk tasks. However, WMSDs are still common among human teleoperators and operators physically interacting with robots in industries. To address these issues, we propose ergonomically intelligent systems for teleoperation and physical human-robot interactions that cover postural estimation, assessment, and optimization. First, we explore estimating the 3D posture of the human using the interacting robot as the sensor. Later we combine it with the OpenPose markerless posture estimation method to prove an occlusion-robust algorithm for posture estimation. We model the posture estimation problem as a partially observable dynamical system and use a particle filter to infer the 3d posture from the sensory observations. We perform a human subject study to evaluate our posture estimation approaches. Next, we use deep neural networks to learn differentiable and continuous ergonomic models based on RULA and REBA, popular and well-studied risk assessment tools in the ergonomics community. We evaluate them by assessing the postures in different types of tasks from two datasets of human motions from subject studies. Finally, we introduce a framework for postural optimization, which includes solutions for different types of teleoperation and physical human-robot interaction tasks. Here, we use gradient-free and gradient-based approaches to solve the optimizations, which benefit from our differentiable ergonomic models. We develop a simulated environment for teleoperation and physical human-robot interaction in which a human performs interactive tasks with robots while accepting and applying the suggested postural corrections from the postural optimization algorithm. [Description provided by NIOSH]
• Subjects:
  Mining Occupational Health Safety Vibration
• Keywords:
  Automation Ergonomics Motion Studies MSD Musculoskeletal Disorders Posture Robotics Sensors
• Publisher:
  University of Utah
• Document Type:
  Text
• Funding:
  Grant
• Genre:
  Dissertation
• Place as Subject:
  OSHA Region 8 ; Utah
• CIO:
  National Institute for Occupational Safety and Health (NIOSH)
• Topic:
  Workplace Safety & Health
• Location:
  North America
• Pages in Document:
  1-81
• NIOSHTIC Number:
  nn:20068737
• Citation:
  Salt Lake City, UT: University of Utah, 2022 Dec; :1-81
• Federal Fiscal Year:
  2023
• Performing Organization:
  University of Utah
• Peer Reviewed:
  False
• Start Date:
  20050701
• Source Full Name:
  Ergonomically intelligent teleoperation and physical human-robot interaction
• End Date:
  20280630
• Collection(s):
  National Institute for Occupational Safety and Health
• Main Document Checksum:
  urn:sha-512:91f189577660ee49c23ef5c370a6d92391f6e8854b93a3a404b3a36356c9e9459ac30e6439c7338b9a4162e5e3a793b48250ea0872408eddd1172710c3e89a98
• Download URL:
  https://stacks.cdc.gov/view/cdc/207323/cdc_207323_DS1.pdf
• File Type:
  [PDF - 17.36 MB ]