An event-based adaptive tactic coordination in tele-operating human-machine interactions using Hybrid Q-learning

Internet based tele-operation systems utilize the ubiquitous connectivity and low cost bandwidth offered by the Internet to send commands and receive friendly feedback for tele-operating remote systems. In many scenarios for efficiency and task completion, the tactic coordination of human-machine and multi-robots are required. For this to be feasible there are several hurdles to be crossed including Internet type delays, uncertainties in the environment and uncertainties in the object manipulated. This paper proposes a general event-based adaptive tele-operation control mechanism that integrates machine learning to deal with such uncertainties. The mechanism was tested by experiments in tele-operation soccer robots system powered by the Hybrid Q-learning and the experimental results that confirmed this mechanism are presented.