Behavior Control System Design Of NAO Robot

With the enlargement of their application fields, more and more intelligent robots have come into people's life. In order to promote the convenient communication between humans and robots and let people enjoy better service, creating of a better human-machine interactive pattern has become one of the important directions in robot research fields. Early interactions of Human-machine are mainly with button and keyboard, but with the development of intelligent robot technology, voice and vision become the main way of human-machine interaction.The design of behavior control system in this paper is addressed based on the NAO robot, which is a research platform produced by Aldebaran Robotics. Two human-machine interaction modes, voice interaction and visual interaction, are realized respectively to control the robot.In voice interaction mode, the input speech is identified with the NAO robot voice recognition module firstly. Then the recognized texts are searched in the database. With the designed voice interaction process, the robot performs the corresponding actions or outputs the voice according to the speech command obtained. Microsoft Speech SDK 5.1 is used in this paper to recognize the word from the voice. A database with three hundred poems of Song Dynasty is constructed. All the operations on the database are realized with Python, which can achieve the results of the query from the database and read it out through the speech synthesis system.In visual interaction mode, target objects are searched and recognized by the camera of the NAO robot. A scheme to identify and localize the position target objects is designed with NAO Mark in this paper. The distance and convert coordinate are calculated according to the acquired NAO Mark information. Furthermore, the pose of NAO Mark relative to the coordinate system of robot is obtained.Finally, motion planning of the NAO robot is studied in order to grab the target object according to the voice command and the information acquired from the visual recognition and localization. Based on that, a behavior control system with both voice and visual interaction are realized, which has more convenient and effective interactions between robot and human. In motion control, a vision-based servo-control system is proposed in this paper. The system can acquire the position of the target object by the camera and plan the arm's motion of robot, which forms a closed-loop control system with visual feedback.