Title: Wearable Sensing and Intervention Strategies for Improving the Biomechanical Safety of Workers

 

Date: Monday, July 6th, 2026

Time: 2:30PM to 4:00PM ET

Location: MRDC 4211 or via zoom (meeting id: 966 6470 1064)

 

Ryan Casey

Robotics Ph.D. Student

Woodruff School of Mechanical Engineering

Georgia Institute of Technology

 

Committee:

Dr. Anirban Mazumdar (advisor) -- Woodruff School of Mechanical Engineering

Dr. Aaron Young (advisor) -- Woodruff School of Mechanical Engineering

Dr. Gregory Sawicki -- Woodruff School of Mechanical Engineering

Dr. Mohsen Moghaddam -- H. Milton Stewart School of Industrial and Systems Engineering

Dr. Jason Wheeler -- Sandia National Labs

 

Abstract:

Manual laborers face an elevated incidence of chronic overuse injury, yet existing protocols for assessing and mitigating this risk remain limited by an inability to characterize injury risk practically and a lack of validated intervention strategies. This thesis proposal examines wearable sensing and intervention systems for reducing joint contact forces (JCFs), the internal mechanical loads most closely tied to musculoskeletal injury. In aim 1, we develop a wearable sensing system combining IMUs and pressure insoles with deep learning models to estimate lower-body biomechanics in real time. We demonstrate that this system outperforms analytical methods across joint angles and moments, and report on ongoing work to refine model training specifically for knee joint loading estimation. In aim 2, we evaluate the effects of four exosuits on knee and lumbar JCFs during simulated occupational lifting tasks, using EMG-informed musculoskeletal modeling that explicitly accounts for exosuit assistance as external forces. We find that both passive and active exosuits can reduce loading at their targeted joints, and that the design and assistance profile of devices heavily impact their performance. We also find that exosuits produce measurable but inconsistent effects at non-targeted joints, motivating the importance of multi-joint analysis. In aim 3, we propose a real-time, task-agnostic biofeedback system that conveys estimated knee joint loads directly to users via augmented reality, and describe a training paradigm designed to guide users toward lifting strategies that reduce knee loading without increasing lumbar loading. Together, these studies advance the development of practical, wearable technologies for characterizing and reducing injury risk in physically demanding occupations.