Development of a Voice-controlled Intelligent Wheelchair System using Raspberry Pi

摘要

This paper proposed a voice recognition-based intelligent wheelchair system for physically disabled people who unable to control the wheelchair by their upper and lower limbs. This development employs voice command to controls the movement of the wheelchair in different directions. The android device is used as microphone to be connected to the Google Assistant prior to data processing by the Raspberry Pi. The Raspberry Pi will then command the servo motors to act accordingly. This system offers automatic obstacle detection via the use of an infrared sensor which aids the user to apply momentary brake button upon the obstacle detection. The proposed development has been evaluated in terms of movement accuracy and system responsiveness upon the commands given, in both control mode i.e., voice controller and joystick controller. For the movement accuracy, the results indicate that the minimum accuracies are 90% and 95% for the voice recognition controller and joystick controller, respectively. For the system responsiveness, the system response in less than 1.2 seconds for the voice recognition controller and less than 0.65 seconds for the joystick controller. With this noticeable performance, the intelligent wheelchair enables users with limited control over their limbs to maneuver the wheelchair using voice commands.