Research On Voice-based Control Of Mobile Robot

Speech recognition is an important branch in the field of the intelligent robot research. Many kinds of robots have become very important in human life today. With the widely used of electronic computer and artificial intelligent machine, People found that the best way of information communication between human and robot is the language exchange. And speech is the expression of acoustic. The dream idea in human mind is to talk with robot without any block. This paper makes a study that the theory and technology of controlling mobile robot by speech. The main research achievements are as followings.Firstly, the development of speech recognition technology is introduced. The present situation and research methods of speech recognition in our country and abroad are elaborated and summarized, advantages and defaults of each method are also analyzed, and all this make an important basis for the research in the paper.Secondly, the speech signal is analysised, including generation model, pretreatment, add window frame points, endpoint detection and feature extraction. And the signal is changed to digital signal in the time domain field.Thirdly, the DTW and HMM algorithm in speech recognition field are introduced. The template matching method of speech recognition is used by DTW algorithm. The algorithm uses dynamic programming method to solve that the feature parameters have long range. It mainly used in which the speecher are not changed and isolated words domain. According to the change of phonetic pronunciation, HMM algorithm calculates the biggest probability to realize the speech recognition. It mainly used in the domain of speaker-independent and small vocabulary recognition. Two experiments based on different algorithm of speech recognition are designed. The experimental results show that, the DTW algorithm mainly used in which the speecher are not changed and isolated words domain; the HMM algorithm mainly used in the domain of speaker-independent and small vocabulary recognition.Finally, the system architecture of MT-ARM mobile robot in laboratory was introduced. The software based on HMM algorithm was designed to control the robot and the mechanical arm of four degrees through human speech. When the robot received speech signals, it had to analysis whether the speech was control command. After the speech recognition, the robot would implement relevant control operation. The feasibility of the proposed strategy was verified by the experimental results.