Controlling a Robot Using Small Database Speech Recognition

Tanasak Phanprasit


Control the movement of the robot in order to obtain maximum efficiency must be considered response functions and the highest level of security as set designer. This paper presents results of testing industrial robots by using small database speech recognition. The small database was created by applying the Fast Fourier Transforms (FFT) to convert speech signal in time domain to speech signal in frequency domain. Then, the speech signal in the frequency domain (data size) was reduced by using the Principal Component Analysis (PCA). The commands were represented by Eigen value and the Eigen vector. The ten speakers consist of 5 males and 5 females. Each speaker was required to utter all seven commands. The results showed the accuracy of the speech recognition, in case the commands came from the ten speakers, was 67.00 percent. Otherwise, the accuracy was 61.29 percent.


Speech Recognition, Fast Fourier, Transform, Principal Component Analysis

Full Text:



Hong Liu, Xiaofei Li, “A Selection Method of Speech Vocabulary for Human-Robot Speech Interaction,” Systems Man and Cybernetics (SMC), IEEE Internation Conference on, pp. 2243-2248, 2010.

Sylvain Calinon, Julien Epiney and Aude Billard, “A Humanoid-Robot Drawing Human Portraits,” The 5th IEEE-RAS International Conference on Humanoid Tsukuba, Japan, Dec. 5-5, pp. 161-166, 2005.

Raji Sukumar. A, Firoz Shah. A, Sarin sukumar. A, Babu. P., “Key-Word Based Query Recognition In a Speech By Using Artificial Neural Network,” Second International Conference Computation Intelligence, Communication System and Networks, 2010.

Jan-Yee Lee, Jeih-weih Hung, “Exploiting Principal Component Analysis in modulation spectrum enhancement for robust speech recognition,” IEEE Trans on audio, 2011 Eighth International Conference on Fuzzy Systems and Knowledge Discovery (FSKD), vol. no.3, pp. 1947-1951, 2011.