Guiding the Human Along the Path Using Dog Robot Based on Harness-Hand Interaction Inspired by Tractor-Trailer Model

Robot guide dogs can serve as substitutes for real guide dogs, which are limited in availability, to ensure the safety of blind or visually impaired (BVI) individuals. For this to be effective, the system must accurately monitor human motion and correct any mistakes in real time, which is challenging due to the need for precise robot positioning while considering human input. This research aims to enhance guidance precision by integrating the human into the control loop using a tractor-trailer model. The system tracks the human's position and pose in real time to correct deviations, with its performance tested in simulated obstacle avoidance scenarios. The implementation effectively corrects lateral errors. However, managing angular errors remains a challenge, as these tend to increase due to the dynamic and unpredictable nature of human motion. This study contributes to the advancement of robotic guide dogs in two main ways: first, by establishing a reliable range for human pathways considering disruptive factors, and second, by assessing the potential improvement in movement precision through continuous monitoring of the human's position.