Teleoperation and Beyond for Assistive Humanoid Robots

Abstract

In this review, we explore how teleoperation could potentially be applied to the management of humanoid robots, with an emphasis on humanoid robots that are used in assistive roles, including clinical therapies. Since there are very few examples of the remote operation of a full humanoid, the review emphasizes technologies that are potentially relevant to teleoperation of humanoids. Teleoperation of humanoid robots faces many of the same practical challenges associated with (a) traditional teleoperation, including latency and telepresence, and (b) teleoperating manipulators and other robots with high degrees of freedom. Teleoperation systems for humanoid robots must also address unique challenges triggered by strong emotional and social responses to humanoids-responses of both the operator and any humans who may interact with the humanoid. These challenges trigger new opportunities for redefining teleoperation to include scripting, programming by demonstration, speech production, and Wizard of Oz interaction. The challenges also provide opportunities to specialize existing modes of interaction and unique opportunities to develop humanoid-specific forms of teleoperation, such as controlling humanoids via exoskeleton-based or inertial sensors. We include in this review a survey of enabling technologies, a taxonomy of practical uses, and a list of emerging themes and approaches to teleoperating humanoid robots.

Bridging the Gap Between Supervised Autonomy and Teleoperation

“…The lack of ability to operate fully autonomously lead to the emergence of approaches that aim at blending human and robot intelligence to ease human-robot interaction on different levels of autonomy (see [8,Tab. 5.2]).…”

“…Their main idea is to ease human-robot interaction by integrating autonomous features supporting the user. While, supervised autonomy allows the operator to initiate tasks on a high level of abstraction that are carried out by the robot autonomously [8], shared control provides support for the user who controls the robot via continuous input on a Human-Robot Interface (HRI) [8], e.g. by means of safequarding.…”

Programming Robots to Express Emotions: Interaction Paradigms, Communication Modalities, and Context

“…R OBOTIC systems are becoming increasingly prevalent in entertainment (e.g., as tour guides [1], [2] and actors in theatrical plays [3]), health care and assistive robotics (e.g., [4], [5]), search and rescue (e.g., [6], [7]), education, military operations, and other domains. In these systems, a robot interacts with end users and other people in the environment.…”

A knowledge-driven shared autonomy human-robot interface for tablet computers

“…The touchscreen of the tablet computer is used to display feedback from the robot to the operator providing the information used for gaining common ground with the robot and situational awareness in the remote environment [6]. An often utilized approach uses the major part of the screen for showing the video stream of a camera mounted on the robot [7][8] optionally augmented with additional sensor [9] or status information [10][11] [12].…”

“…However, real world environments often demand a combination of both [6]. Muszynski et al [11] proposed a tablet computer based HRI allowing the operator to adjust the autonomy of the robot by switching between commanding of basic movements, semiautonomous capabilities or high level tasks.…”