User Evaluation of an Interactive Learning Framework for Single-Arm and Dual-Arm Robots

Abstract

Abstract. Social robots are expected to adapt to their users and, like their human counterparts, learn from the interaction. In our previous work, we proposed an interactive learning framework that enables a user to intervene and modify a segment of the robot arm trajectory. The framework uses gesture teleoperation and reinforcement learning to learn new motions. In the current work, we compared the user experience with the proposed framework implemented on the single-arm and dual-arm Barrett's 7-DOF WAM robots equipped with a Microsoft Kinect camera for user tracking and gesture recognition. User performance and workload were measured in a series of trials with two groups of 6 participants using two robot settings in different order for counterbalancing. The experimental results showed that, for the same task, users required less time and produced shorter robot trajectories with the single-arm robot than with the dual-arm robot. The results also showed that the users who performed the task with the single-arm robot first experienced considerably less workload in performing the task with the dual-arm robot while achieving a higher task success rate in a shorter time.

Design of an Autonomous Trash-Picking Service Robot Focussed on Human-Robot Interaction

“…Research in physical human-robot interaction (p-HRI) has drastically increased because of more robots being introduced for varied use cases (Jevtić et al, 2016). However, only a few works have gone into incorporating hardware design elements in service robots, such as in the development of the humanoid service robot R1 (Lehmann et al, 2016;Parmiggiani et al, 2017).…”

Robot motion adaptation through user intervention and reinforcement learning

“…In the previous work [10], we assessed the usability of an interactive learning framework based on user intervention and reinforcement learning (RL), to allow users to modify an unfitted segment of a robot's trajectory. The motion adjustment was done using hands' gestures to guide the robot along a corrective path.…”

Bimanual robot skills: MP encoding, dimensionality reduction and reinforcement learning

“…While latest trends in AI are also using deep learning to solve several complex problems, the amount of data in real environments is too limited to apply those deep learning approaches. As shown in our work [23] [133], few samples can suffice to learn simple tasks from scratch, but there's still a long way to make robots capable of correctly adapting motion and compliance to any new situation.…”

“…We then went on building a global friction model for the WAM robot, allowing for a more precise dynamics model that could be used all throughout the robot's workspace [21] (see video B.2 in Appendix B). We also applied DREPS to human environments, using hybrid methods for gathering data, by visual imitation or randomly-generated samples in [22,23]. Video B.4 in Appendix B show this hybrid approach.…”

“…6.8 and, using RL, the robot had to adapt the trajectory to an S-shaped motion that would not knock down any of the bottles. Within this problem, we learned the task using a compliant controller as defined in Part I of this thesis in three different manners: Through human guidance [22,23,133], with REPS [121] and with DREPS.…”