A Voice Command Detection system for controlling Movement of SCOUT Robot

Human interaction with robots can be achieved in different ways, depending on the type of robots that humans intend to communicate with. According to the increasing social uses of robots in recent years, this article has been focused on human-robot interaction via speech. For this purpose, a voice command detection system is designed to control the SCOUT robot. The voice command detection system is implemented based on Hidden Markov model. We create a non-native speech database for desired commands. This data base includes Ready, Forward, Backward, Left, Right, Stop, Circle, Faster, Slower, and Go commands. Each command has been uttered by 20 Persian speakers (10 males and 10 females). The voice commands are detected using the HMM-based system. Then, the detected command is sent to the robot through the Bluetooth module. The results of evaluating the system show an acceptable performance of the human-robot interaction.