Learning to Collaborate From Simulation for Robot-Assisted Dressing

Clegg, Alexander; Erickson, Zackory; Grady, Patrick; Turk, Greg; Kemp, Charles C.; Liu, C. Karen

doi:10.1109/lra.2020.2972852

Cited by 49 publications

(49 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…With the accessibility of physics simulation for physical human-robot interaction (see Assistive Gym [92]), simulated capacitance measurements can serve as a feature for safely learning intelligent robot controllers across a diverse array of human body shapes and sizes. As a step towards simulated capacitive sensors, Clegg et al [84] introduced a simulated multidimensional distance sensor for measuring the shortest Cartesian distance to the human body within a simulated robot-assisted dressing environment.…”

Section: Discussionmentioning

confidence: 99%

See 1 more Smart Citation

Characterizing Multidimensional Capacitive Servoing for Physical Human-Robot Interaction

Erickson¹,

Clever²,

Gangaram³

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

Towards the goal of robots performing robust and intelligent physical interactions with people, it is crucial that robots are able to accurately sense the human body, follow trajectories around the body, and track human motion. This study introduces a capacitive servoing control scheme that allows a robot to sense and navigate around human limbs during close physical interactions. Capacitive servoing leverages temporal measurements from a multi-electrode capacitive sensor array mounted on a robot's end effector to estimate the relative position and orientation (pose) of a nearby human limb. Capacitive servoing then uses these human pose estimates from a datadriven pose estimator within a feedback control loop in order to maneuver the robot's end effector around the surface of a human limb. We provide a design overview of capacitive sensors for human-robot interaction and then investigate the performance and generalization of capacitive servoing through an experiment with 12 human participants. The results indicate that multidimensional capacitive servoing enables a robot's end effector to move proximally or distally along human limbs while adapting to human pose. Using a cross-validation experiment, results further show that capacitive servoing generalizes well across people with different body size.

show abstract

Section: Discussionmentioning

confidence: 99%

“…In healthcare and assistive robotics settings, sensing approaches that rely on force sensors [78], pressure mats [79], [80], [81], capacitive sensors [82], [83], [84], and feature extraction from occluded depth images [85], [70] have been applied to recover visually occluded human pose.…”

Section: Human Pose Estimation For Physical Assistancementioning

confidence: 99%

Characterizing Multidimensional Capacitive Servoing for Physical Human-Robot Interaction

Erickson¹,

Clever²,

Gangaram³

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…Recently the authors of [25] address the problem from safety perspective and proposed a motion planning strategy that theoretically guarantees safety under the uncertainty in human dynamic models. Zhang et al, uses a hybrid force/position control with simple planning for dressing [26], while [27] use deep reinforcement learning (DRL) to simultaneously train human and robot control policies as separate neural networks using physics simulations. Although DRL yields satisfactory dressing policies, applying DRL in a real world setting is very difficult, especially if the task involves a human.…”

Section: Robotic Experimentsmentioning

confidence: 99%

Learning Task-Parameterized Skills from Few Demonstrations

Zhu¹,

Gienger²,

Kober³

2022

Preprint

View full text Add to dashboard Cite

Moving away from repetitive tasks, robots nowadays demand versatile skills that adapt to different situations. Taskparameterized learning improves the generalization of motion policies by encoding relevant contextual information in the task parameters, hence enabling flexible task executions. However, training such a policy often requires collecting multiple demonstrations in different situations. To comprehensively create different situations is non-trivial thus renders the method less applicable to real-world problems. Therefore, training with fewer demonstrations/situations is desirable. This paper presents a novel concept to augment the original training dataset with synthetic data for policy improvements, thus allows learning task-parameterized skills with few demonstrations. Videos of the experiments are available at https://sites.google.com/view/ tp-gmm-from-few-demos/

show abstract

“…Accordingly, the TO used all ellipsoids for hand position and arm configuration, as well as ellipsoids for hand velocity, when ensuring safety constraints. We have considered lifting this assumption in future work by drawing insights from prior art [9,50,20,6,19].…”

Section: H M 4 Z P I S Z E E G Y C K Z U R W S K Z a Y A J N P W Y T W 9 S N 8 N 3 T L T L O Z K Z D X P V Z / H U C E N I F Z C A K C U mentioning

confidence: 99%

Provably Safe and Efficient Motion Planning with Uncertain Human Dynamics

Shen¹,

Figueroa²,

Shah³

et al. 2021

Robotics: Science and Systems XVII

View full text Add to dashboard Cite

Ensuring human safety without unnecessarily impacting task efficiency during human-robot interactive manipulation tasks is a critical challenge. In this work, we formally define human physical safety as collision avoidance or safe impact in the event of a collision. We developed a motion planner that theoretically guarantees safety, with a high probability, under the uncertainty in human dynamic models. Our two-pronged definition of safety is able to unlock the planner's potential in finding efficient plans even when collision avoidance is nearly impossible. The improved efficiency is empirically demonstrated in both a simulated goal-reaching domain and a real-world robot-assisted dressing domain. We provide a unified view of two approaches to safe human-robot interaction: human-aware motion planners that use predictive human models and reactive controllers that compliantly handle collisions.

show abstract

Learning to Collaborate From Simulation for Robot-Assisted Dressing

Cited by 49 publications

References 31 publications

Characterizing Multidimensional Capacitive Servoing for Physical Human-Robot Interaction

Characterizing Multidimensional Capacitive Servoing for Physical Human-Robot Interaction

Learning Task-Parameterized Skills from Few Demonstrations

Provably Safe and Efficient Motion Planning with Uncertain Human Dynamics

Contact Info

Product

Resources

About