Visdom: Smart guide robot for visually impaired people

Lee, Zhen Ting (2025) Visdom: Smart guide robot for visually impaired people. Final Year Project, UTAR.

Preview

PDF Download (2711Kb) | Preview

Abstract

This report details the development of an autonomous guide robot system for visually impaired individuals using the ROS 2 framework and the Wheeltec R550 Robot platform. Key features include the SMAC Hybrid A* algorithm for global path planning, the Regulated Pure Pursuit Controller for local trajectory execution and dynamic obstacle avoidance, and gmapping for SLAM functionality. The system architecture integrates ROS 2 on a Raspberry Pi, with TCP/IP connectivity enabling remote operation. An Android mobile application, developed using Java and the java.net.Socket library, provides an intuitive and accessible user interface for seamless interaction with the robot. The app incorporates TalkBack accessibility support, enabling visually impaired users to navigate the interface using screen reader feedback. Once a destination is selected, the robot guides the user using a combination of voice prompts (e.g., "Heading to kitchen", "Be careful to your left") and physical feedback through a mounted guiding stick. The project has demonstrated successful results in mapping and navigation within static indoor environments. Core functionalities such as path planning, autonomous movement, voice feedback, and app-to-robot communication have been thoroughly tested and optimized. The integration of the mobile interface and real-time voice prompting significantly enhances user accessibility and confidence. This system represents a significant advancement in assistive technology, offering improved mobility and independence for visually impaired individuals. The project showcases the practical application of autonomous navigation technologies in real-world scenarios, highlighting the potential of robotics and AI in improving quality of life for people with visual impairments.

Actions (login required)