Embedded IoT-Based Smart Assistive Glove System For Real-Time Gesture Recognition And Emergency Alerts

Communication is a daily struggle for people with speech and hearing impairments whenever the person across from them does not know sign language, and that struggle turns serious in an emergency, when getting help depends on being understood right away. Vision-based recognition, wearable sensing, and sensor-fusion methods have each moved sign lan-guage interpretation forward, but every one runs into the same problems: heavy computation, poor portability, slow response, or privacy loss. This paper works through these gesture recognition approaches, weighing what each does well against where it falls short, and looks at how embedded machine learning and edge computing are changing what a wearable device can do. One gap keeps showing up across the literature: almost none of these communication-focused systems build in real emergency support, and the few that try rarely fit how a nonverbal or hearing-impaired person would want to trigger one. Building on that gap, this paper proposes EchoHand, an IoT-enabled glove pairing flex sensors, a six-axis IMU, and an ESP32 microcontroller to recognize gestures and produce speech or text output, while reusing that pipeline to send a location-aware alert when needed. It closes by covering the practical hurdles this design faces and where it goes next.