Development Algorithm and Assessing the Efficacy of Voice Control Robotic Prosthetic Hand

Abstract

This research investigates the development and evaluation of a voice-controlled robotic prosthetic hand, offering a potentially more intuitive and user-friendly interface for individuals with upper limb differences. The system utilizes a Raspberry Pi as the central processing unit, leveraging the Python Speech Recognition library and the Google Cloud Speech API for speech-to-text conversion and command recognition. Five servo motors, controlled via a PCA9685 driver board, actuate the prosthetic hand’s fingers, mimicking essential grasping and individual finger movements. The performance of the system was assessed through rigorous testing with three participants, focusing on metrics such as word recognition accuracy, command success rate, and overall system latency. Results demonstrated high recognition accuracy, exceeding 97% across all participants, confirming the effectiveness of the chosen speech recognition engine. Command success rates were also consistently high, indicating reliable translation of spoken commands into the intended hand movements. However, the “grip” command presented challenges due to phonetic similarities with other words, highlighting the need for further optimization in speech recognition. Analysis of the system latency revealed that audio capture and processing time on the Raspberry Pi was the dominant contributor to overall delay, suggesting potential benefits from exploring local speech recognition methods. The servo motor performance was consistently fast and accurate, confirming the viability of the mechanical design and control strategy. This research successfully demonstrates the feasibility of voice control for robotic prosthetic hands, providing a foundation for future development and highlighting the importance of addressing pronunciation variability, optimizing latency, and incorporating user feedback for improved usability.