Robert G. Radwin
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Robert G. Radwin has a total of 176 co-authors. Here are the top co-authors by number of shared publications:
Yu Hen Hu
University of Wisconsin-Madison
13 shared publications- Bilge Mutlu
Department of Computer Sciences, University of Wisconsin-Madison, Madison, Wisconsin, USA
9 shared publications - Michael Gleicher
Department of Computer Sciences, University of Wisconsin-Madison, Madison, WI, USA
9 shared publications David P. Azari
From the Department of Industrial and Systems Engineering (D.P.A., R.G.R.); Department of Surgery (S.R.P.Q., C.C.G.), Clinical Sciences Center; Department of Urology (B.V.L.); and Duane H. and Dorothy M. Bluemke Professor in the College of Engineering (R.G.R.), University of Wisconsin-Madison, Madison, WI; Department of Surgery (L.L.F.), Penn Medicine - University of Pennsylvania Health System, Philadelphia, PA; City of Hope National Comprehensive Cancer Center (B.L.M), Duarte, CA
8 shared publicationsMary E. Sesto
Department of Medicine, University of Wisconsin-Madison, Madison, WI, USA
6 shared publications- Michael Hagenow
University of Wisconsin-Madison, Madison, USA
6 shared publications
Latest Publications
- 15 June 2026
- Published by SAGE Publications in Human Factors: The Journal of the Human Factors and Ergonomics Society
- Vol. 68 (12) , 1561-1579
Modeling Human Expertise in a Sanding Task
Objective: Sanding strategies were identified that correlated with better performance and less physical stress. Background: Sanding requires manual skill because of the dexterity and perceptual needs to perform the job well, and it involves exposure to physical stresses. Methods: Force and tool kinematics were measured using an instrumented platform and motion capture for 20 participants while performing a timed paint removal sanding task in the laboratory. Performance was based on the quantity and uniformity of the paint removed and measured using image processing. The ACGIH Threshold Limit Value (TLV) for Hand Activity Level Peak Force Index was used to evaluate physical stress. Results: Linear regression and machine learning algorithms revealed that high performers (above the 75th percentile) moved the sander more, while they varied but applied greater force compared to the low performers (below the 25th percentile). Some high performers could maintain a TLV below the acceptable limit by pausing more and applying less force but performing more frequent exertions. Conclusions: More force, greater variations of force, and speed were related to better sanding performance, while less force and more frequent exertions and pauses helped reduce physical stress. Applications: Strategies identified in this study may be useful for training operators in sanding tasks for performance and safety.
- 15 June 2026
- Published by SAGE Publications in Human Factors: The Journal of the Human Factors and Ergonomics Society
- Vol. 68 (12) , 1561-1579
Modeling Human Expertise in a Sanding Task
Objective: Sanding strategies were identified that correlated with better performance and less physical stress. Background: Sanding requires manual skill because of the dexterity and perceptual needs to perform the job well, and it involves exposure to physical stresses. Methods: Force and tool kinematics were measured using an instrumented platform and motion capture for 20 participants while performing a timed paint removal sanding task in the laboratory. Performance was based on the quantity and uniformity of the paint removed and measured using image processing. The ACGIH Threshold Limit Value (TLV) for Hand Activity Level Peak Force Index was used to evaluate physical stress. Results: Linear regression and machine learning algorithms revealed that high performers (above the 75th percentile) moved the sander more, while they varied but applied greater force compared to the low performers (below the 25th percentile). Some high performers could maintain a TLV below the acceptable limit by pausing more and applying less force but performing more frequent exertions. Conclusions: More force, greater variations of force, and speed were related to better sanding performance, while less force and more frequent exertions and pauses helped reduce physical stress. Applications: Strategies identified in this study may be useful for training operators in sanding tasks for performance and safety.
- Open Access
- 27 October 2025
- Published by Springer Nature in Machine Vision and Applications
- Vol. 36 (6) , 1-14
Towards video-based injury risk assessment: predicting lifting loads from body pose trajectories
Manual material handling tasks, such as lifting and lowering, are ubiquitous across industry sectors. Overexertion during these tasks is among the leading causes of workplace injuries. Previous studies have shown that lifting load is a key factor in determining the risk of injury. However, existing methods for measuring the lifting load often rely on manual measurements, sensor fusion, or other techniques that are difficult to scale in practice. In this study, we present a vision-based approach to automatically predict lifting load by analyzing human body pose trajectories extracted from video alone. Specifically, our method employs person detection, visual tracking, and human body pose estimation to extract pose trajectories and their kinematic features, which are then used to train a Transformer model for load prediction. To evaluate our method, we conducted a human subjects study of 19 participants performing various lifting and lowering tasks with varying postures. Our method achieved an average accuracy of 74.8% to distinguish between light vs. heavy objects, and an average accuracy of 50.8% to identify three levels of lifting loads (light, medium, heavy) across lifting and lowering tasks. These results demonstrate a first step towards computer vision based solutions for automatic, noninvasive, scalable injury risk assessment for manual material handling tasks.
- 27 October 2025
- Published by Springer Nature in Machine Vision and Applications
- Vol. 36 (6) , 1-14
Open AccessTowards video-based injury risk assessment: predicting lifting loads from body pose trajectories
Manual material handling tasks, such as lifting and lowering, are ubiquitous across industry sectors. Overexertion during these tasks is among the leading causes of workplace injuries. Previous studies have shown that lifting load is a key factor in determining the risk of injury. However, existing methods for measuring the lifting load often rely on manual measurements, sensor fusion, or other techniques that are difficult to scale in practice. In this study, we present a vision-based approach to automatically predict lifting load by analyzing human body pose trajectories extracted from video alone. Specifically, our method employs person detection, visual tracking, and human body pose estimation to extract pose trajectories and their kinematic features, which are then used to train a Transformer model for load prediction. To evaluate our method, we conducted a human subjects study of 19 participants performing various lifting and lowering tasks with varying postures. Our method achieved an average accuracy of 74.8% to distinguish between light vs. heavy objects, and an average accuracy of 50.8% to identify three levels of lifting loads (light, medium, heavy) across lifting and lowering tasks. These results demonstrate a first step towards computer vision based solutions for automatic, noninvasive, scalable injury risk assessment for manual material handling tasks.
- 2 October 2025
- Published by Taylor & Francis in IISE Transactions on Occupational Ergonomics and Human Factors
- Vol. 13 (4) , 217-219
Introduction to the Festschrift for Professor Thomas J. Armstrong
- 2 October 2025
- Published by Taylor & Francis in IISE Transactions on Occupational Ergonomics and Human Factors
- Vol. 13 (4) , 217-219
Introduction to the Festschrift for Professor Thomas J. Armstrong
- 26 September 2025
- Published by Taylor & Francis in IISE Transactions on Occupational Ergonomics and Human Factors
- Vol. 13 (4) , 290-300
Ergonomics Analysis for a Simulation Approach to Human-Robot Collaborative Task Allocation
Occupational ApplicationsWe used an algorithmic approach to assign tasks among a human worker and a collaborative robot assistant. When optimized on force and time as a measure for ergonomics cost, the algorithm assigned physically demanding aspects of the job to the robot, such as applying forces, while the human worker was assigned more dexterous tasks, such as the alignment of parts. Using this approach, operator stress was reduced while the cycle time for the job decreased. This method may be useful for better integrating collaborative robots into jobs requiring ergonomics interventions.
- 26 September 2025
- Published by Taylor & Francis in IISE Transactions on Occupational Ergonomics and Human Factors
- Vol. 13 (4) , 290-300
Ergonomics Analysis for a Simulation Approach to Human-Robot Collaborative Task Allocation
Occupational ApplicationsWe used an algorithmic approach to assign tasks among a human worker and a collaborative robot assistant. When optimized on force and time as a measure for ergonomics cost, the algorithm assigned physically demanding aspects of the job to the robot, such as applying forces, while the human worker was assigned more dexterous tasks, such as the alignment of parts. Using this approach, operator stress was reduced while the cycle time for the job decreased. This method may be useful for better integrating collaborative robots into jobs requiring ergonomics interventions.
- 25 February 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Human-Machine Systems
- Vol. 55 (2) , 309-314
A Single-Camera Method for Estimating Lift Asymmetry Angles Using Deep Learning Computer Vision Algorithms
A computer vision (CV) method to automatically measure the revised NIOSH lifting equation asymmetry angle (A) from a single camera is described and tested. A laboratory study involving ten participants performing various lifts was used to estimate A in comparison to ground truth joint coordinates obtained using 3-D motion capture (MoCap). To address challenges, such as obstructed views and limitations in camera placement in real-world scenarios, the CV method utilized video-derived coordinates from a selected set of landmarks. A 2-D pose estimator (HR-Net) detected landmark coordinates in each video frame, and a 3-D algorithm (VideoPose3D) estimated the depth of each 2-D landmark by analyzing its trajectories. The mean absolute precision error for the CV method, compared to MoCap measurements using the same subset of landmarks for estimating A, was 6.25° (SD = 10.19°, N = 360). The mean absolute accuracy error of the CV method, compared against conventional MoCap landmark markers was 9.45° (SD = 14.01°, N = 360).
- 25 February 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Human-Machine Systems
- Vol. 55 (2) , 309-314
A Single-Camera Method for Estimating Lift Asymmetry Angles Using Deep Learning Computer Vision Algorithms
A computer vision (CV) method to automatically measure the revised NIOSH lifting equation asymmetry angle (A) from a single camera is described and tested. A laboratory study involving ten participants performing various lifts was used to estimate A in comparison to ground truth joint coordinates obtained using 3-D motion capture (MoCap). To address challenges, such as obstructed views and limitations in camera placement in real-world scenarios, the CV method utilized video-derived coordinates from a selected set of landmarks. A 2-D pose estimator (HR-Net) detected landmark coordinates in each video frame, and a 3-D algorithm (VideoPose3D) estimated the depth of each 2-D landmark by analyzing its trajectories. The mean absolute precision error for the CV method, compared to MoCap measurements using the same subset of landmarks for estimating A, was 6.25° (SD = 10.19°, N = 360). The mean absolute accuracy error of the CV method, compared against conventional MoCap landmark markers was 9.45° (SD = 14.01°, N = 360).
- 8 August 2024
- Published by SAGE Publications in Human Factors: The Journal of the Human Factors and Ergonomics Society
- Vol. 67 (3) , 246-263
Human–Robot Collaboration With a Corrective Shared Controlled Robot in a Sanding Task
Objective: Physical and cognitive workloads and performance were studied for a corrective shared control (CSC) human–robot collaborative (HRC) sanding task. Background: Manual sanding is physically demanding. Collaborative robots (cobots) can potentially reduce physical stress, but fully autonomous implementation has been particularly challenging due to skill, task variability, and robot limitations. CSC is an HRC method where the robot operates semi-autonomously while the human provides real-time corrections. Methods: Twenty laboratory participants removed paint using an orbital sander, both manually and with a CSC robot. A fully automated robot was also tested. Results: The CSC robot improved subjective discomfort compared to manual sanding in the upper arm by 29.5%, lower arm by 32%, hand by 36.5%, front of the shoulder by 24%, and back of the shoulder by 17.5%. Muscle fatigue measured using EMG, was observed in the medial deltoid and flexor carpi radialis for the manual condition. The composite cognitive workload on the NASA-TLX increased by 14.3% for manual sanding due to high physical demand and effort, while mental demand was 14% greater for the CSC robot. Digital imaging showed that the CSC robot outperformed the automated condition by 7.16% for uniformity, 4.96% for quantity, and 6.06% in total. Conclusions: In this example, we found that human skills and techniques were integral to sanding and can be successfully incorporated into HRC systems. Humans performed the task using the CSC robot with less fatigue and discomfort. Applications: The results can influence implementation of future HRC systems in manufacturing environments.
- 8 August 2024
- Published by SAGE Publications in Human Factors: The Journal of the Human Factors and Ergonomics Society
- Vol. 67 (3) , 246-263
Human–Robot Collaboration With a Corrective Shared Controlled Robot in a Sanding Task
Objective: Physical and cognitive workloads and performance were studied for a corrective shared control (CSC) human–robot collaborative (HRC) sanding task. Background: Manual sanding is physically demanding. Collaborative robots (cobots) can potentially reduce physical stress, but fully autonomous implementation has been particularly challenging due to skill, task variability, and robot limitations. CSC is an HRC method where the robot operates semi-autonomously while the human provides real-time corrections. Methods: Twenty laboratory participants removed paint using an orbital sander, both manually and with a CSC robot. A fully automated robot was also tested. Results: The CSC robot improved subjective discomfort compared to manual sanding in the upper arm by 29.5%, lower arm by 32%, hand by 36.5%, front of the shoulder by 24%, and back of the shoulder by 17.5%. Muscle fatigue measured using EMG, was observed in the medial deltoid and flexor carpi radialis for the manual condition. The composite cognitive workload on the NASA-TLX increased by 14.3% for manual sanding due to high physical demand and effort, while mental demand was 14% greater for the CSC robot. Digital imaging showed that the CSC robot outperformed the automated condition by 7.16% for uniformity, 4.96% for quantity, and 6.06% in total. Conclusions: In this example, we found that human skills and techniques were integral to sanding and can be successfully incorporated into HRC systems. Humans performed the task using the CSC robot with less fatigue and discomfort. Applications: The results can influence implementation of future HRC systems in manufacturing environments.
- Open Access
- 11 March 2024
- Published by Association for Computing Machinery (ACM)
- p. 231-239
A System for Human-Robot Teaming through End-User Programming and Shared Autonomy
Many industrial tasks-such as sanding, installing fasteners, and wire harnessing-are difficult to automate due to task complexity and variability. We instead investigate deploying robots in an assistive role for these tasks, where the robot assumes the physical task burden and the skilled worker provides both the high-level task planning and low-level feedback necessary to effectively complete the task. In this article, we describe the development of a system for flexible human-robot teaming that combines state-of-the-art methods in end-user programming and shared autonomy and its implementation in sanding applications. We demonstrate the use of the system in two types of sanding tasks, situated in aircraft manufacturing, that highlight two potential workflows within the human-robot teaming setup. We conclude by discussing challenges and opportunities in human-robot teaming identified during the development, application, and demonstration of our system.
- 11 March 2024
- Published by Association for Computing Machinery (ACM)
- p. 231-239
Open AccessA System for Human-Robot Teaming through End-User Programming and Shared Autonomy
Many industrial tasks-such as sanding, installing fasteners, and wire harnessing-are difficult to automate due to task complexity and variability. We instead investigate deploying robots in an assistive role for these tasks, where the robot assumes the physical task burden and the skilled worker provides both the high-level task planning and low-level feedback necessary to effectively complete the task. In this article, we describe the development of a system for flexible human-robot teaming that combines state-of-the-art methods in end-user programming and shared autonomy and its implementation in sanding applications. We demonstrate the use of the system in two types of sanding tasks, situated in aircraft manufacturing, that highlight two potential workflows within the human-robot teaming setup. We conclude by discussing challenges and opportunities in human-robot teaming identified during the development, application, and demonstration of our system.
- 25 October 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Robotics and Automation Letters
- Vol. 8 (12) , 8335-8342
Coordinated Multi-Robot Shared Autonomy Based on Scheduling and Demonstrations
Shared autonomy methods, where a human operator and a robot arm work together, have enabled robots to complete a range of complex and highly variable tasks. Existing work primarily focuses on one human sharing autonomy with a single robot. By contrast, in this letter we present an approach for multi-robot shared autonomy that enables one operator to provide real-time corrections across two coordinated robots completing the same task in parallel. Sharing autonomy with multiple robots presents fundamental challenges. The human can only correct one robot at a time, and without coordination, the human may be left idle for long periods of time. Accordingly, we develop an approach that aligns the robot's learned motions to best utilize the human's expertise. Our key idea is to leverage Learning from Demonstration (LfD) and time warping to schedule the motions of the robots based on when they may require assistance. Our method uses variability in operator demonstrations to identify the types of corrections an operator might apply during shared autonomy, leverages flexibility in how quickly the task was performed in demonstrations to aid in scheduling, and iteratively estimates the likelihood of when corrections may be needed to ensure that only one robot is completing an action requiring assistance. Through a preliminary study, we show that our method can decrease the scheduled time spent sanding by iteratively estimating the times when each robot could need assistance and generating an optimized schedule that allows the operator to provide corrections to each robot during these times.
- 25 October 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Robotics and Automation Letters
- Vol. 8 (12) , 8335-8342
Coordinated Multi-Robot Shared Autonomy Based on Scheduling and Demonstrations
Shared autonomy methods, where a human operator and a robot arm work together, have enabled robots to complete a range of complex and highly variable tasks. Existing work primarily focuses on one human sharing autonomy with a single robot. By contrast, in this letter we present an approach for multi-robot shared autonomy that enables one operator to provide real-time corrections across two coordinated robots completing the same task in parallel. Sharing autonomy with multiple robots presents fundamental challenges. The human can only correct one robot at a time, and without coordination, the human may be left idle for long periods of time. Accordingly, we develop an approach that aligns the robot's learned motions to best utilize the human's expertise. Our key idea is to leverage Learning from Demonstration (LfD) and time warping to schedule the motions of the robots based on when they may require assistance. Our method uses variability in operator demonstrations to identify the types of corrections an operator might apply during shared autonomy, leverages flexibility in how quickly the task was performed in demonstrations to aid in scheduling, and iteratively estimates the likelihood of when corrections may be needed to ensure that only one robot is completing an action requiring assistance. Through a preliminary study, we show that our method can decrease the scheduled time spent sanding by iteratively estimating the times when each robot could need assistance and generating an optimized schedule that allows the operator to provide corrections to each robot during these times.
- 1 September 2023
- Published by SAGE Publications in Proceedings of the Human Factors and Ergonomics Society Annual Meeting
- Vol. 67 (1) , 845-846
Human-Robot Collaboration in a Sanding Task
Robots can help reduce physical stress in manual sanding, but implementation has been challenging due to the skill required, task variability, and robot limitations. In shared control (SC) a robot operates semi-autonomously while the human provides real-time corrections. A laboratory study tested the effectiveness of SC for removing paint using an orbital sander. A robot autonomously sanded the surface, and twenty participants sanded manually and in collaboration with a robot using SC. Subjective discomfort for manual sanding was greater than SC in the upper arm by 28.5%, the lower arm by 29%, the hand by 38%, and the shoulder by 42%. Muscle fatigue, measured using EMG, was 22.4% greater for the manual condition. Cognitive workload measured using the NASA TLX was 14.25% more for manual sanding. Digital imaging showed that SC outperformed the fully autonomous robot by 10.75% for uniformity, by 4.96% for quantity, and by 6.06% for average performance.
- 1 September 2023
- Published by SAGE Publications in Proceedings of the Human Factors and Ergonomics Society Annual Meeting
- Vol. 67 (1) , 845-846
Human-Robot Collaboration in a Sanding Task
Robots can help reduce physical stress in manual sanding, but implementation has been challenging due to the skill required, task variability, and robot limitations. In shared control (SC) a robot operates semi-autonomously while the human provides real-time corrections. A laboratory study tested the effectiveness of SC for removing paint using an orbital sander. A robot autonomously sanded the surface, and twenty participants sanded manually and in collaboration with a robot using SC. Subjective discomfort for manual sanding was greater than SC in the upper arm by 28.5%, the lower arm by 29%, the hand by 38%, and the shoulder by 42%. Muscle fatigue, measured using EMG, was 22.4% greater for the manual condition. Cognitive workload measured using the NASA TLX was 14.25% more for manual sanding. Digital imaging showed that SC outperformed the fully autonomous robot by 10.75% for uniformity, by 4.96% for quantity, and by 6.06% for average performance.
- Open Access
Handheld Haptic Device with Coupled Bidirectional Input
Handheld kinesthetic haptic interfaces can provide greater mobility and richer tactile information as compared to traditional grounded devices. In this paper, we introduce a new handheld haptic interface which takes input using bidirectional coupled finger flexion. We present the device design motivation and design details and experimentally evaluate its performance in terms of transparency and rendering bandwidth using a handheld prototype device. In addition, we assess the device's functional performance through a user study comparing the proposed device to a commonly used grounded input device in a set of targeting and tracking tasks.
Open AccessHandheld Haptic Device with Coupled Bidirectional Input
Handheld kinesthetic haptic interfaces can provide greater mobility and richer tactile information as compared to traditional grounded devices. In this paper, we introduce a new handheld haptic interface which takes input using bidirectional coupled finger flexion. We present the device design motivation and design details and experimentally evaluate its performance in terms of transparency and rendering bandwidth using a handheld prototype device. In addition, we assess the device's functional performance through a user study comparing the proposed device to a commonly used grounded input device in a set of targeting and tracking tasks.