Research And Design Of Voice Control Systems Of Painting Robots

In the new era of increasing automation,the spraying production line industry has been widely used,and coating robots have been added to the spraying production line.In the process of controlling and using the painting robot,the traditional method is through "dumb" manual control,which has been unable to meet people's needs.Therefore,human-specific language communication,the most easy-to-understand and most frequent communication method,will surely become the most popular way in the human-computer interaction market in the future.The research of robots plus voice control functions will play a vital role in the development of "humanized" robots in the future.This article has conducted some researches on the problems involved in the voice control system of painting robots.The specific work is as follows:1.Analyze the preprocessing of the speech recognition system,the speech framing,windowing function,and endpoint detection.Reminiscent of the smooth lowpass characteristics of the speech spectrum,this paper chooses the Hamming window function as the windowing function.The dual-threshold endpoint detection method adopts different parameters to simulate experiments.According to the characteristics of low interference noise and long time duration,a new dual-threshold endpoint detection method is selected to use its short term with better noise resistance.The elapsed threshold rate is used as a parameter.2.Based on the analysis of three mainstream feature parameter extraction methods,based on the accuracy of speech recognition,using linear predictive cepstrum(LPCC)and Mel frequency cepstrum(MFCC)feature parameters,this paper proposes a New voice signal feature extraction method.The relative error rate of the other three feature parameter extractions is compared through experiments,and it is verified that the method reduces the loss of accuracy.3.Researched the traditional hidden Markov model(HMM model)three problem solving algorithms,combined with the characteristics of the speech signal,segmented through the correlation degree between frames,removed the self-transition arc,and proposed a new The training algorithm of the HMM model,through simulation comparison,obtains smaller training error,and the training convergence speed is faster,which reflects the superiority of the improved method.4.Realize the design and programming of the speech recognition system,use the Windows API to obtain the recording audio method,call the audio recording related interface functions,complete the design and production of the entire system and the expected functions,and realize the painting robot adds the voice recognition function Imagine making it more intelligent.