Forming an Emotional Response in a Robot

Problem statement: effective communication between robots and people requires taking into account the role of emotions in the mechanism of behavior control and joint practical activities. Recognizing human emotions and the robot demonstrating signs of an emotional response that are appropriate to the situation are important elements of successful interaction. Emotions have a wide range of manifestations, and trying to technically implement a robot to demonstrate signs of the entire spectrum of emotions is difficult. Purpose of research: analysis of ways to reduce the complexity of the task of demonstrating by a robot signs of an effective emotional response appropriate to the situation. In this case, a necessary condition should be the ability to unambiguously recognize the class of demonstrated emotions. Assessment of the possibility of choosing the type of emotional response based on the development of decision-making models with unclear initial data about the mixed emotional state of the person participating in communication. Results: The real requirements for the emotional manifestations of a robot during working interaction with a person, including within the framework of social robotics, are considered. Analysis of job descriptions that prescribe employee behavior when performing these jobs, expert opinions show that it is possible to significantly narrow the range of response emotions demonstrated by the robot, sufficient to complete the assigned tasks, reduce them to several clearly distinguishable classes, with technically accessible hardware implementation. Limiting the number of response emotions demonstrated by the robot allows the use of a simplified analysis of human emotions that generate the robot’s response, the use of discrete models, and forming of data for decision-making using fuzzy logic. Practical significance: A mechanism for the formation of an emotional response of a robot participating in working interaction with a person is proposed. It allows you to limit and simplify the multidimensional meanings of feature describing emotions, and to use methods of logical–linguistic analysis to make decisions, reduces the likelihood of a robot’s emotional reaction not matching the situation.