Assistive mobility technologies have become increasingly important in improving the quality of life for individuals with disabilities. However, individuals with dual sensory loss, including visual and hearing impairments, continue to face major challenges in safe navigation and environmental awareness. Traditional mobility aids such as white canes and guide dogs provide only limited support because they cannot identify dynamic obstacles or hazardous environmental conditions in real time. This research proposes “SenseStep: AI-Integrated Smart Assistive Footwear for Dual Sensory Loss Individuals,” an intelligent wearable assistive system that combines artificial intelligence, embedded sensors, and wireless communication technologies to improve mobility assistance and user safety. The proposed system integrates ultrasonic sensors, flame sensors, and water sensors to detect nearby obstacles and environmental hazards. A camera module combined with a YOLO-based deep learning model performs real-time object recognition and distance estimation. The collected sensor and vision data are processed by a microcontroller to generate structured vibration-based feedback and optional Braille output, enabling communication without relying on visual or auditory cues. The system also includes a wireless communication module for sending emergency alerts and location updates to caregivers during hazardous situations. In addition, piezoelectric energy harvesting technology is integrated to improve battery efficiency by converting walking pressure into electrical energy. Experimental testing demonstrated an overall detection accuracy of 92%, with obstacle detection accuracy of 94%, hazard detection accuracy of 90%, and object recognition accuracy of 93%. The proposed system significantly improves safety, accessibility, and independence for individuals with dual sensory impairments by providing an intelligent, reliable, and energy-efficient assistive mobility solution

Keywords— Artificial Intelligence; Smart Footwear; YOLO; Assistive Technology; Embedded Systems; Object Detection