Optimization model for wireless charging and power saving of smart canes for the visually impaired based on DRL

Abstract

With the growing demand for intelligent mobility aids, smart canes for visually impaired individuals require efficient energy management and reliable charging solutions. This study presents an optimization model for wireless charging and power-saving in smart canes, leveraging deep reinforcement learning (DRL). Unlike conventional charging strategies that rely on static scheduling, our model dynamically optimizes charging decisions using a Deep Q-Network (DQN)-based algorithm, considering real-time environmental factors and user behavior. Additionally, an adaptive energy-saving strategy is proposed to regulate the operation of key functional modules—voice guidance, ultrasonic detection, alarms, time announcements, and flashlight warnings—based on contextual needs. Experimental evaluations demonstrate significant improvements in charging efficiency, battery longevity, and user experience compared to traditional methods. By integrating reinforcement learning with intelligent energy management, this research provides an innovative and practical approach to enhancing smart cane functionality, promoting safer and more autonomous navigation for visually impaired individuals.