Designing Persuasive Robotics: Using Emotion Recognition in Humanoid Robots to Persuade Humans

This thesis explores how socially interactive humanoid robots can influence human behavior and decision-making through the act of persuasion. Computational systems, combined with human-like physical form and function, enable such humanoid robots to act in a way to persuade humans. Currently, there is very little understanding of the persuasive potential of such machines. As personal robots become more commercially accessible, a better understanding of the mechanisms and the capabilities of their ability to influence is becoming increasingly important. This thesis proposes some guiding principles by which to qualify persuasion done by humanoid robots. A pilot study was designed in which the Nao robotic platform was used to solicit research participants for donations for a community food pantry. The study tests some verbal and nonverbal behavioral variables known to affect persuasiveness in humans and measures the effect of these variables in human-robot interaction. The results of this pilot study indicate that factors such as robot-gender, interpersonal distance, vocal variety, and emotion recognition may have an impact on the interaction between humans and robots, therefore, influencing the design of sociable robots.

This is a pilot study built upon the work done by Mikey Siegel and Cynthia Breazeal from MIT. An extensive study will be done in the future and results will be published.