Assisted Teleoperation for a Human-Swarm Interaction System

Abstract

Abstract The trend in human-controlled robotic systems is to move away from the current many-to-one paradigm, where multiple operators are forced to control a single system to a one-to-many mode of operation in which a single operator can control a large number of autonomous vehicles. For this transition to be successful, it is essential that we clearly understand how humans should interact with a large number of robots, this is precisely the problem of a growing field of research called Human-Swarm Interaction (HSI). In this article, the training of a recurrent neural network (LSTM) is proposed to model the behavior of the human operator in the task of rendezvous performed by swarms, in order to develop an assisted teleoperation system. The system is tested in the presence of humans and in the presence of faults in the communications.