Deceptive Actions and Robot Collision Avoidance

Abstract

The problem of collision avoidance is extensively studied in contemporary robotics. Efficiency and effectiveness of multirobot task cooperation and safety of human-robot interactions depend critically on the quality of the solution of the problem of collision avoidance. Classical path planning and obstacle avoiding approaches can not guarantee a sufficiently high quality of the solution of the problem of multirobot and human collision avoidance. So, to solve the problem with a sufficiently high quality, some special algorithms and approaches are needed. In this paper, we consider the notion of deception in context of the problem. We consider the possibility of generating and using deception to manipulate another robot. In particular, we consider two robots that solve their own tasks. In addition, we consider two robots one of which solves its own task and the second robot simulates the behavior of a pedestrian. We study the possibility of using fraud to improve the robot’s own performance and to improve safety of collision avoidance.