Explainable Autonomous Mobile Robots: Interface and Socially Aware Learning

With the growing utilization of mobile robots in urban environments, one can expect more interaction of mobile robots with people in shared spaces. It is important that these robots navigate among people in a socially acceptable manner and construct proper interactions with their surroundings. Robot behavior relies on Artificial Intelligence (AI) to intercept the environment and command the actuators based on the perception. Inadequate explainability of autonomous robots' decisions can cause unnecessary mental and physical risks. Prevention and mitigation of the hazards are possible with proper robot to human interfaces. This paper presents systems and methods designed for self-explaining mobile robots in common areas using projected information. Collected data about the reactions of people to the robot with and without self-explainable features is analyzed in different operational domains on university campus, such as makerplace, hallways, cafeterias, and study spaces. Our initial analysis based on the robot's task completion performance and questionnaires from people demonstrates noticeable enhancements in acceptance of the robot.