Intention modeling of Web user for networked mobile robot with Bayesian inference

Teleoperated networked robot often has unpredictable behaviors due to uncertain time delay in data transmission over Internet. This paper investigates a novel approach to compensate the uncertain time delay with the autonomy of mobile robot. The user´s intention to operate and control the networked robot is incrementally inferred based on Bayesian techniques so that task of the robot defined by the user´s intention can be recognized. Consequently, the robot can autonomously execute this task without frequent interaction with the user, which decreases data transmission to a great extent. So long the robot has a correct inference on task, the efficiency of the whole system can be improved to a remarkable extent. Experimental results show the validity and feasibility of the proposed method.