Sign in Subscribe
Machine learning

Robotics Researchers Develop Guide Robot for Visually Impaired Individuals

Strategic Insights

— 2 min read

Researchers at National Taiwan Ocean University have developed an indoor guide robot designed to assist visually impaired individuals by physically guiding them to a specified destination. The robot employs a camera, object detection, and a closed-loop control system to manage its arm movements and navigate around obstacles. A Dueling Double Deep Q Network (D3QN) enhanced with a genetic algorithm is used for path planning, and a 3D point cloud map is generated to determine the start and end points accurately.

Key Takeaways:

  • The guide robot is designed to physically assist visually impaired individuals by holding their hand with a robotic arm and guiding them to a specified destination.
  • The robot employs a camera combined with object detection to identify the human hand and a closed-loop control system to manage its arm movements.
  • A Dueling Double Deep Q Network (D3QN) enhanced with a genetic algorithm is used for path planning to ensure the robot safely guides blind individuals to their destinations without collisions.
  • The robot utilizes a depth camera alongside fuzzy logic to control its wheels and navigate around obstacles.
  • A 3D point cloud map is generated to determine the start and end points accurately, enabling the robot to plan its path effectively.
  • The guide robot is a joint research project between Chen-Hsien Yang, Department of Communications, Navigation & Control Engineering, National Taiwan Ocean University, and Jih-Gau Juang, National Taiwan Ocean University.
  • The development of this robot aims to enhance the independence and mobility of visually impaired individuals.

Statistics:

  • The research was published in the journal Machines in 2025, volume 13, issue 7, page 560 (Machines - http://www.mdpi.com/journal/machines).
  • The guide robot is designed to navigate indoor spaces, where obstacles and collisions pose a significant risk to visually impaired individuals.
  • The D3QN algorithm is used to compensate for dynamic obstacles and adjust the robot's path to ensure safe navigation (10.3390/machines13070560).
  • The genetic algorithm enhances the D3QN algorithm by introducing variables that improve the robot's ability to plan its path effectively (10.3390/machines13070560).
  • The research aims to improve the independence and mobility of visually impaired individuals, with a focus on indoor navigation and object avoidance.

Sources:

  • NewsRx. Researchers from National Taiwan Ocean University Publish Findings in Robotics (Guide Robot Based on Image Processing and Path Planning). Life Science Weekly. August 12, 2025; p 6078.
  • Guide Robot Based on Image Processing and Path Planning. Machines, 2025, 13(7):560. (Machines - http://www.mdpi.com/journal/machines).
  • MDPI AG. Machines. (n.d.). Retrieved from http://www.mdpi.com/journal/machines.

Read more