Methods for Expressing Robot Intent for Human–Robot Collaboration in Shared Workspaces

Abstract

Human–robot collaboration is becoming increasingly common in factories around the world; accordingly, we need to improve the interaction experiences between humans and robots working in these spaces. In this article, we report on a user study that investigated methods for providing information to a person about a robot’s intent to move when working together in a shared workspace through signals provided by the robot. In this case, the workspace was the surface of a tabletop. Our study tested the effectiveness of three motion-based and three light-based intent signals as well as the overall level of comfort participants felt while working with the robot to sort colored blocks on the tabletop. Although not significant, our findings suggest that the light signal located closest to the workspace—an LED bracelet located closest to the robot’s end effector—was the most noticeable and least confusing to participants. These findings can be leveraged to support human–robot collaborations in shared spaces.

SLOT-V: Supervised Learning of Observer Models for Legible Robot Motion Planning in Manipulation

“…One of the first major approaches to modelling legibility has been proposed by Dragan et al [6], [7] and has sparked extensive follow up [10], [11], [12], [13], [14], [15], [16]. The work assumes that humans expect robots to move efficiently and that we can model this expectation using a cost function over trajectories (the observer model).…”

A Review on Communicative Mechanisms of External HMIs in Human-Technology Interaction

“…The same is likely to apply to autonomous vehicles and other autonomous systems that require complex interactions with people, where people are familiar with the normal motions associated with specific contexts. [41] This has been tested by e.g. Arntz et al [38], who tested various interfaces, including a text-based interface and mounting simple OK / not OK lights on the robot arm, and found that these basic systems led to mostly increased positivity towards the robot's motion, but did not necessarily increase trust or efficiency.…”

“…Unfortunately, research on the interaction between Baxter's facial expressions and human actors is quite scarce although some sources do exist on the matter. Lemasurier et al [41] investigated and compared the use of light emitters and motion clues of a Baxter robot and found that it is primarily light signals in close proximity to the end effector that is most efficiently caught by a human actor. Their variables included two instances where the robot's screen was used, one instance where the eye gaze was the clue and one instance where the actual pan of the "head" or screen was the factor.…”

“…It should also be pointed out that all instances of the independent variables were significantly better than the control in noticeability, indicating that using either movement or light signals are both reasonable approaches. While Lemasurier et al's [41] study does provide valuable information (more than what has been covered here) it should be kept in mind that it is a study simply looking at whether a robot limb is to be moved or not and has very little to say on the trajectory or goal of said movement. To be able to work with movement trajectories of robot limbs, either additional light sources to indicate direction, speed, etc., or a deeper look at the nature of movements, would be required.…”

“…LEGIBLE MOTION While adding light fixtures as a communicative feature on either autonomous cars, AGVs, or collaborative robots could provide some clues that an action is to take place, they seem to be limited in their communicative capabilities when interacting with humans. They might even complicate and change the interaction as seen in the AGV example [31], or they might simply be insufficient in providing the human with information about the upcoming motion [41]. Understanding and predicting, not only that the robot is about to move, but also what the goal of the action is, requires better and possibly more 'natural' communicative features of the robot.…”