Echo Guidance: Voice-Activated Application for Blind with Smart Assistive Stick Using Machine Learning and IoT

Navigating the world without sight presents profound challenges for millions of visually impaired individuals. Echo Guidance includes a pioneering application, addresses these obstacles by facilitating independence and mobility through voice-driven interaction and an IoT smart assistive device for obstacle detection by utilizing Ultrasonic, GPS and GSM modules powered by Arduino. This study explores the design and evaluation of Echo Guidance, highlighting its utilization of advanced algorithms like RGB888 conversion, matrix transformation, and Optical Character Recognition (OCR) for real-time object recognition. The application boasts a voice-centric interface tailored to the needs of the visually impaired, seamlessly integrating day-to-day functionalities such as weather forecasting, reminders, and a calculator accessible through voice commands. Central to Echo Guidance is its innovative object detection and recognition approach, employing the MobileNetObjDetector class powered by TensorFlow Lite for swift and accurate object detection. The OverlayView class visually represents detected objects, enhancing spatial awareness. Additionally, the integration of IoT devices, such as the smart stick with ultrasonic sensors for obstacle detection and head-level obstacle detection, coupled with SOS functionality in case of any emergency using GPS and GSM modules, further enhances Echo Guidance's capabilities. This research underscores Echo Guidance's transformative potential in improving the quality of life for visually impaired individuals, representing a significant advancement in assistive technology. Prioritizing accessibility and user experience, this paper empowers individuals worldwide with newfound independence and confidence in navigating their surroundings.