A Smart Wearable Device for Stuttering Detection and Intervention using IoT and Machine Learning Technologies

Abstract

This research paper presents the development and evaluation of an IoT-based wearable assistive device designed for stuttering monitoring and feedback, integrating machine learning algorithms for enhanced accuracy and real-time functionality. The device incorporates a variety of hardware and software components, including sensors, microcontroller units, and communication protocols, to capture speech patterns and movement accurately. Machine learning algorithms, such as Support Vector Machines (SVM) and Convolutional Neural Networks (CNN), are employed for real-time analysis of speech patterns and detection of stuttering episodes. The selection criteria for machine learning models, training and testing procedures, and performance evaluation metrics are discussed in detail. The wearable device prototype underwent rigorous testing and validation, demonstrating high accuracy, sensitivity, and specificity in distinguishing between stuttered and fluent speech patterns. User feedback and usability evaluations highlighted the device's ergonomic design, intuitive interface, and real-time feedback capabilities, positioning it as a promising tool for improving stuttering therapy outcomes. Comparative analysis with existing solutions further underscored the device's superior performance and potential for clinical adoption. The implications of the findings for stuttering therapy, limitations, future work, and potential for clinical adoption are discussed, emphasizing the device's contributions to personalized and effective interventions in stuttering therapy. Overall, this research contributes to the advancement of assistive technologies in speech therapy and highlights the potential of IoT-based wearable devices integrated with machine learning algorithms for improving the quality of life for individuals with stuttering disorders.