Comfortable Following Evaluation Index System for Service Robots in Dynamic Environments

Abstract

Human following in a dynamic environment is one of the hotspots in the field of service robot research. The characteristics of the traditional following method cannot meet people’s requirements for a new generation of service robots, and the human-robot coexist model has gradually become the mainstream. Tri-Co robots not only need to complete tasks, but also need to consider maintaining human physical safety and psychological comfort while performing following tasks. This paper proposes an evaluation index system of robot comfortable following. Three experimental scenes of office area, corridor, and lobby are designed to verify the effectiveness and comfort of the robot’s following behavior. As a result, the evaluation shows that the method proposed in this paper improves the psychological comfort of the followed human and other pedestrians compared with the traditional following method. Service robots can show humanlike social behaviors, which increases the degree of robots communicating with human and will be more acceptable to humans.