Human-robot Interactions And Emotions

Human-Robot Interaction: The Impact of Robot-displayed Emotions on Performance in Team Situations

By Graduate Student

Contents

Overview 2

Introduction 2

Emotion 3

Appearance 4

Summary 4

Identified Research Questions 4

Anticipated Outcomes 4

References 5

Overview

I explore the impact of robot-displayed emotions on performance in Human-Robot Interaction (HRI) team situations. Robots have been utilized in team situations for decades. In the mid 1980s a breakthrough with work behavior-based, autonomous robot technology allowed for robots to use distributed sense-response loops to generate appropriate responses to specific external stimuli, which are robust to changes in the environment. The next big step occurred with hybrid architectures that allowed simultaneous reactive behaviors and high-level cognitive reasoning, which are fundamental to successful HRI (Encyclopedia Brittanica) [1].

As we step through the doorway from the past of intelligent robotic invention to the future of HRI innovation decisions about whether or not to integrate emotions into work-related situations must be deliberated. If the decision to merge is chosen, then specific emotions must be determined as the preferred options for clear and immediate communication as directly related to the task or work environment. Humans convey emotions as a means of communication. If a picture speaks a thousand words, then a change in expression compounds those thousand words exponentially. Because there are so many variations of basic expressions, it is best to stick with a limited few to avoid confusion in communication.

Introduction

Robots function as tools, assistants and peers in team situations, which means that they are controlled by a human counterpart to some degree depending upon the demand of the task and the design of the robot. The focus of this paper is on peer-to-peer human-robot interaction in specific work-related team situations where the robot is subordinate, has local autonomy within a prescribed set of routines. and is vital to the team's success. Specifically described are the interactions between humans and robots where the robots have displayed emotional characteristics similar to those found in humans, and the impact of those emotions in a team environment. This particular robot has the standard appearance of an autonomous work-robot, adapted from the Disney/Pixar motion picture Wall-E, see Fig. 1 with facial features approximately in the same areas as with human features. This robot's features have the ability to move to create basic emotional expressions such as movable eyebrow flaps in a neutral position to indicate that all is well, a nod of the articulated head to indicate agreement, a shake of the articulated head to indicate disagreement and exaggerated jerking movements by articulated arms to indicate alarm. Individually articulated fingers were also considered for the purpose of this study.

Fig. 1 (Michael, 2009) [2]

Emotion

The best social competence model for emotion communication lies in the robot's ability to respond to human behavior without the robot initiating intentional social contact. In this way the robot relies “...on humans to attribute mental states and emotions to the robot.” Fong et al. [3] stated that a survey by Khan [4], revealed that people's attitudes are strongly influenced by science fiction. From the survey two significant findings were: (1) a robot with machine-like appearance, serious personality, and round shape is preferred, and; (2) verbal communication using a human-like voice is highly desired.

According to a recent article published on GIZMODO.com, a project called, “Feelix Growing” has an international team of psychologists, neuroscientists and roboticists working towards building robots with human emotions that adapt to our needs (Ramirez, 2007) [5]. These robots would react to happy or sad auditory stimulus and respond accordingly with celebratory or consolatory emotions. While that is an significant achievement towards social robots, it would not be advantageous in a work environment.

The work environment, while meant to be a place to focus on required and paid activity is not totally devoid of emotion. On the contrary, the work place has a harmony and feel to it that affects the human worker. Having a robot respond in an appropriate, yet subdued, emotional manner would be a welcome addition to the rhythm of the job. There's an old saying that goes: you can laugh at your own jokes as long as you don't laugh louder than your audience. This holds true for robots as well. It would be another step into the 'uncanny valley' for robots to be reacting with robust emotional responses instead of the expected and safe nod, or shake.

The potential to develop trust in a robot human-partner may rely on simple, yet consistent conveyances of subdued emotion responses. In addition to facilitating communication through emotion responses a learning feature would be of enormous benefit if the robot could through artificial intelligence continually update its emotion-cue reaction response with every human-robot interaction.

There is the danger that in creating an emotional bond between the human-worker and his/her robot partner it may impair the work relationship. There is some evidence that suggests that soldiers working in high stakes situation with his robot-counterpart, like mine diffusion, develop emotional bonds with the robots, naming them and grieving when the robot sustains serious damage (Groom, 2008) [6]. There may not be an opportunity for robots in extreme work situations to have emotion-cued responses that will not inadvertently put humans in danger. Humans will and do respond to emotion displayed by the robot. Computers as Social Actors

...