A prototype wheelchair controlled via voice commands in Turkish.
Built with Raspberry Pi, Vosk speech recognition, and GPIO motor control.
The Voice-Activated Wheelchair project aimed to provide an affordable assistive mobility solution for users with limited physical control. By leveraging speech recognition and microcontroller-based control, the system enabled hands-free navigation using simple commands.
The prototype successfully mapped real-time speech commands to motor actions, providing intuitive and responsive control. Latency was minimized through offline recognition, making the system robust without internet dependency.
Speech recognition accuracy varied with background noise. To improve performance, I applied noise filtering and optimized the microphone placement. For safety, I integrated an ultrasonic sensor-based cutoff to stop the wheelchair if an obstacle was detected within range.