Home Service Robot Control System Based On Speech Signals Recognition

This paper studies the process of speech signals recognition:mainly by the speech signals filtering, sampling, quantization, windowing, endpoint detection, feature extraction, model training and threshold comparison and the realization of the model simulation algorithms through matlab. While through matlab GUI design techniques achieving a speech recognition interface. By building the five degrees of freedom robot arm Arduino and ASR MO8-A speech recognition module, Realization of voice control robot movement.The speech signal after filtering, sampling and quantization, obtaining discrete digital signal. Pre-emphasis, pre-emphasis is to filter out low frequency interference, enhance the high-frequency component of the input signal. Sub-frame makes the original signal into a section of the section, Corresponds to the original signal plus a rectangular window in the time domain. Signal is multiplied by a rectangular window in time domain, Equivalent to Signal spectrum with the Fourier transform of the rectangular window convolution in the Frequency domain.Double threshold endpoint detection algorithm to realize the endpoint detection of the speech signal by short-time energy and zero crossing rate.After double threshold endpoint detection algorithm detect of the speech signal endpoint,the Mel frequency cepstrum coefficient and the first order differential Mel frequency cepstrum coefficients obtain the characteristic parameters of the speech signal. Also proposed to improve speech signal feature extraction algorithm. Calculation procedure of linear prediction cepstral based on wavelet transform and Mel frequency cepstral calculation steps based on wavelet transform. Finally, theDWTL、△DWTL、DWTMand△DWTMcombination as the characteristic parameter of the speech signal.And then through the hidden Markov model which, the forward backward algorithm, Viterbi algorithm, Baum-welch algorithm train the model.Meanwhile by the Matlab GUI design and the callback function achieve the speech recognition simulation interface.On the basis of the theory of the speech recognition, combined with Arduino robot arms. Five degrees of freedom robot arm by rotating the common coordinate transformation algorithm achieve forward kinematics and inverse kinematics problem solving of robot arm. The forward kinematics problem is known by robot every joint variable to solve end effector pose, The inverse kinematics problem is based on robot end effector positions andorientations calculate every joint angle. Then use ASR MO8-A speech recognition module,32-channel servo control board, Arduino Atmegal2560panel, Five degrees of freedom manipulator.voice control robot arm.