With the increase of social concern about the disables and elderly people, their participation in social activities is demanded. In this view, motorized wheelchair system, one of the mobile robot, is necessary for giving them better mobility and for saving them a considerable physical. However, it still proves to be difficult or impossible to use for some handicapped person. Therefore, the teaching method with speech command is necessary to a handicapped person without hands or a non-expert to generate the path.
This paper presents the design of intelligent mobile robot system using a real time speech recognition in implementation of motorized wheelchair. The proposed mobile robot system is composed of four separated module, which are main control module, speech recognition module, servo motor driving module and sensor module.
In main control module with microprocessor(80C196KC), one part of the artificial intelligences, fuzzy logic, was applied to the proposed intelligent control system. In order to improve the non-linear characteristic which depend on an user's weight and variable environment, encoder attached to the servo motors was used for feedback control.
Also, in this paper, we design and implement a speaker independent recognition system using TI's DSP(TMS320C32). This system uses Hidden Markov Models for the designated command vocabularies to control a mobile robot, and it has postprocessed by RBFNet(Radical Basis Function Neural Network) to distinguish some fuzzy word command well. As the spectral analysis method, we use a MFCC(Mel Frequency Cepstral Coefficient) to extract the features of the voice.
The proposed recognition system is tested using 9 words for control of the mobile robot, and the performance of a mobile robot using voice and joystick command is also evaluated.