The need for complementing robot's autonomy behaviour with human reasoning has made robot teleoperation a research topic for more than fifty years. And while most of the research work usually deploys a master-slave architecture for controlling robots (whether using a joystick and/or desktop computer), recently Mixed Reality has come into play. This is because it is able to improve the teleoperator's view with less bandwidth consumption than traditional teleoperation architectures. In this paper a new approach for teleoperating a manipulator is presented. The virtual representation of the manipulator is made by using encoder sensors, and also, a camera image is displayed through an Android interface to complement the teleoperator's view. We can show with our results that the real positions of the manipulator and the virtual model are closed to be the same with some differences due to the transmission time delay which is minimum.