Assistive teleoperation of robot arms via automatic time-optimal mode switching

By Laura Herlant, Rachel Holladay, Siddhartha S. Srinivasa. 2016

132 citations recorded by the source corpus.

Assistive robotic arms are increasingly enabling users with upper extremity disabilities to perform activities of daily living on their own. However, the increased capability and dexterity of the arms also makes them harder to control with simple, low-dimensional interfaces like joy-sticks and sip-and-puff interfaces. A common technique to control a high-dimensional system like an arm with a low-dimensional input like a joystick is through switching between multiple control modes. However, our interviews with daily users of the Kinova JACO arm identified mode switching as a key problem, both in terms of time and cognitive load. We further confirmed objectively that mode switching consumes about 17.4% of execution time even for able-bodied users controlling the JACO. Our key insight is that using even a simple model of mode switching, like time optimality, and a simple intervention, like automatically switching modes, significantly improves user satisfaction.

Subfields: Mechanical Engineering, Rehabilitation, Reinforcement Learning, Human-Computer Interaction, Robotics & Control

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Mechanical Engineering

In Rehabilitation

In Reinforcement Learning

In Human-Computer Interaction

In Robotics & Control

Citations in the corpus

LILA: Language-Informed Latent Actions (2021, background): “While the human can manually teleoperate the arm by fully controlling individual “modes”, or separate degrees-of-freedom of the robot’s end-effector, past work has shown this to be unintuitive, slow, and frustrating [3, 6].”
Learning latent actions to control assistive robots (2021, unknown): “End-effector teleoperation is commonly used by assistive robots [22], where the human presses a button to switch modes and control different aspects of the endeffector’s motion.”
Controlling Assistive Robots with Learned Latent Actions (2019, methodology): “with two 2-DoF joysticks, and were given a button to toggle between linear and angular motion [3], [7], [25].”
Learning Visually Guided Latent Actions for Assistive Teleoperation (2021, methodology): “Alongside our assistive approaches, we also include an end-effector control baseline, where users directly teleoperate the robot’s motion by switching control modes (Herlant et al., 2016).”
Shared Autonomy with Learned Latent Actions (2020, background): “This can be extremely taxing for the users [12].”
Shared autonomy via hindsight optimization for teleoperation and teaming (2017, methodology): “Feeding with an assistive arm is difficult (Herlant et al., 2016), and prior work indicates that users subjectively prefer more assistance when the task is difficult even though they have less control (You and Hauser, 2011; Dragan and Srinivasa, 2013b).”
Learning User-Preferred Mappings for Intuitive Robot Control (2020, background): “Other works mitigate modeswitching with model-based optimization [8] or in a datadriven manner [20], but the underlying mode switching still remains discontinuous.”
Kinova Modular Robot Arms for Service Robotics Applications (2017, methodology): “Figure 22 shows a MICO robot at Carnegie Mellon University being used for research in intelligent algorithms to help enable people living with disabilities to perform activities of daily living and instrumental activities of daily living (Herlant et al., 2016).”
Autonomy in Rehabilitation Robotics: An Intersection (2018, unknown): “For example, instead of issuing control commands, the autonomy might instead anticipate and facilitate switches between the control subspaces, to ease the burden of mode switching (135) or to better estimate human intent (136).”