The proposed robot system can be divided into Humanrobot
interaction block and server event block processing.
Human-robot interaction block describes the user behavior
with the robot, which includes the speech control and the touch
control. In this block, human-robot interaction capability is
provided by the RMICS, since the robot only transfers the data
of sensors and displays the result to users. In the speech control
path, we use the scheme of wake-up keyword to enhance the
robustness of voice activity detector (VAD). We adopt the
robot's name as the wake-up keyword. While the user says the
robot's name at first, the robot can respond a beep sound to the
user. This sound means that the user can begin to speak the
speech command within 3 seconds. Otherwise, the wake-up
state of the robot would be canceled; the user should say the
name again to the robot and ready to speak the command. This
wake-up scheme can significantly improve the robustness of
VAD. On the other hand, user can select service by touching
menu in touch panel.