Research And Implementation Of Intelligent Wheelchair Voice Recognition And Control System

With the development of modern science and technology, intelligent robots havebecome a hot research in the field of robotics, robot voice recognition has got widespreadattention by many researchers both at home and abroad in recent years. Intelligentwheelchair is an important research area for HelpAge/Disabled intelligent service robots,add voice recognition technology to a wheelchair, it is very convenient for users to controlwheelchair by voice.The intelligent wheelchair voice recognition and control system was designed, voicerecognition algorithm was studied, main contents are as follows:Firstly, intelligent wheelchair voice-driven control new scheme was designed whichSPCE061A MCU was used as voice information processor and DSP2407A was used asdrive controller, the work principle of system and the function of each hardware circuit wereintroduced.Secondly, the hardware circuit for intelligent wheelchair voice recognition and controlsystem was designed, the hardware circuit include: power supply circuit, motor drivercircuit, the mode switch and joystick circuits, serial communication circuit, the voltagedetection circuit etc, a complete hardware system was built.Then, the software development environment of SPCE061A MCU and TMS320LF24-07A DSP chip, the main program flow chart, subroutine flow chart, interrupt the programflow chart were introduced, and software program was prepared. After software andhardware debugging, the system can realize that the motor can be controlled by voice ormanually, speed values and battery power can be observed through liquid crystal display.The system achieve a better control effect.Finally, voice recognition fuzzy support vector machine algorithm was studied, adetailed design of voice recognition system was given. The double-sphere method wasintroduced to the calculations of fuzzy support vector machine membership function, andisolated word voice recognition system simulation experiments were designed in matlab. Bychoosing a better linear kernel function, experimental results show that fuzzy support vectormachine algorithm in the limited sample, noisy environments can achieve better recognitionresults than DTW algorithm.The intelligent wheelchair voice recognition and control system achieved good results in the experiment, it laid a good foundation for the design of more complex intelligentwheelchair in future. The study of voice recognition based on fuzzy support vector machinesalgorithm provided a technical support for intelligent wheelchair voice recognition in theoutdoor noise environment.