SmartStep: A Wearable Indoor Navigation and Obstacle Detection Shoe for the Visually Impaired Person
Maryam1, Revathy G Krishna2, V K Nandana3, Dr. Meril Cyriac4
1Maryam, Electronics & Computer Engineering, LBSITW
2Revathy G Krishna, Electronics & Computer Engineering, LBSITW
3V K Nandana, Electronics & Computer Engineering, LBSITW
4Dr. Meril Cyriac, HOD, Electronics & Computer Engineering, LBSITW
Abstract - This work aims to develop SmartStep, a hybrid wearable navigation system designed to improve indoor mobility and spatial awareness for visually impaired individuals. We hypothesize that integrating inertial tracking, visual markers, and proximity sensing will significantly enhance navigation accuracy and real-time obstacle detection in complex environments. SmartStep combines Pedestrian Dead Reckoning (PDR) with QR/Aruco marker–based localization to achieve precise indoor positioning, while lightweight TensorFlow Lite object detection models enable continuous environmental awareness. A smartphone functions as the central processing unit, using IMU sensors for step and direction estimation and its camera for marker recognition and object identification. Audio cues delivered through earphones provide navigation guidance and obstacle alerts, supported by a mode-switching mechanism that optimizes power and computational efficiency. The wearable component consists of dual ESP32-based shoe modules, each equipped with front, left, and right ultrasonic sensors that activate vibration motors to deliver immediate tactile feedback. Operating independently without wireless communication, the shoe modules ensure low-latency, real-time obstacle detection. By combining motion tracking, vision-based localization, and proximity sensing, SmartStep offers a modular, cost-effective, and responsive assistive solution that enhances safety, mobility, and independence for visually impaired users in both structured indoor spaces and dynamic environments.
Keywords: Pedestrian Dead Reckoning (PDR), Object Detection, ArUco Markers, Ultrasonic Sensors, ESP32, TensorFlow Lite, Assistive Navigation, Wearable Technology, Indoor Localization, Smart Footwear.
