Analysis and implementation of the PTZ-class facial tracking on humanoid robot

The neck joints are considered as the most fragile structure on the robot skull. However, the facial scenarios can hardly act well without the dexterous servo mechanism such as the PTZ(pen-tilt-zoom) or the harmonically driven mechanism. Structure limitations critically restrict the mechatronic design and dynamic robustness of the intelligent controller of the robot-head module. This paper is aimed to analyze the neck dynamics and develop an image tracking stratagem to perform different facial emotions on the robot skull. A PC-based platform with distributed human-machine interface is applied to supervise the pen-tilt motions and simulate the scenario of the robot head using a modulized infrared CCD. Finally, tracking controlled PTZ (pen-tilt zoom) mechanism is accomplished and computer simulations, facial tracking experiments are successfully made on a self-fabricated humanoid robot.