Aiming at the problems that the traditional speech package information is redundant and the universality and code portability of mobile robot control system are poor

a speech control system for mobile robot based on robot operating system（ ROS） is proposed. The speech control command signal is preprocessed by preemphasis

windowing frame detection

and endpoint detection. The feature parameters are extracted by the Bark wavelet packet transform method based on Fisher ratio and the hidden Markov model（ HMM） algorithm is used for recognition. The above contents are written by C + + language. The software package uses the node function of ROS and the communication mechanism to use the recognition result to control the movement of the mobile robot.The experimental results show that the recognition rate of the system is above 94 %

which can accurately identify the voice control commands of different users. Compared with the mature SDK function

the recognition rate of the system is improved by 1. 6 % and the recognition time is shortened by only 255. 6 ms. The system increases the speed of response while ensuring accuracy.