IoT-Enabled Smart Stick with Multi-Sensor Fusion and LSTM-Based Navigation for Enhanced Mobility

Visual impairment maintains substantial obstacles to independent movement because proper assistance technology should be applied for its resolution. The proposed system outperforms conventional walking sticks because it merges sensors from four categories including ultrasonic, LiDAR and infrared also water detection to improve network perception. The Multi-Sensor Fusion technology allows data sensors to merge intelligently with an effective outcome of real-time decisions and decreased incorrect response rates. Historical movement data goes into an LSTM-based navigation module to forecast safety routes for stopping accidental collisions during use. Using connected IoT technology allows for the implementation of Location Trace functions while ensuring instant emergency SMS messaging with audio feedback delivered through wireless earpieces that produce complete user assistance systems. The adapter filters connected to sensors ensure that the system function reliably in both inside and outside environments. The study establishes three main objectives to achieve: •Develop an advanced assistive mobility solution with enhanced obstacle detection accuracy. •Predictive path planning based on LSTM networks will allow dynamic navigation as part of the research approach. •Ensure real-time location tracking and emergency alert systems for user safety. •Utilize IoT-cloud connectivity for data storage and continuous learning. Evaluation results indicate the developed MSF-LSTM system produces obstacle detection performance of 97.8% surpassing traditional models. The system achieves three main operational advantages through its high false detection prevention along with 2.1-second fast response times and 91.3% better navigation systems. The combination of AI-based assistance with IoT connectivity generates new security measures for autonomous vehicles that benefit users who are vision-impaired. The mobility assistance functions will receive enhancements through terrain modification based on machine-learning and live environment object identification capabilities.