IOT-Based Obstacle Detection and Alert System for the Visually Impaired Person with Voice

Abstract

The system develops an iot based obstacle detection and warning system which improves safety and awareness while enabling visually impaired people to move freely .The system operates through continuous environmental surveillance which uses various sensors that includes distance sensors  and echo sensors and passive infrared motion sensors and flame detection .An ensemble  machine learning system with voting classifier assess real time sensor data to identify four different environmental states  which includes normal and warming  and alarm and voice and collision.The system uses a safety override mechanism based on rules to handle high -risk situations which creates an alert when an obstacle comes with in 10 cm of the machine learning system .

The system performs data preprocessing and feature scaling before classification to enhance prediction accuracy .The system delivers real time multi-channel alert through asynchronous audio notification that use text to speech technology to provide clear voice feedback about obstacle proximity and motion detection and fire hazard .A remote monitoring module send real time alert which include sensor data and threat assessment to a communication platform for ongoing observation and incident documentation the system features a complete training and evaluation system which tests various classifiers together with the final ensemble model  while it automatically creates performance assessment results .

The experimental results establishing that the system successfully identifies threats and send alerts on schedule which proves its effectiveness as an intelligent assistive system for users who are visually impaired