Design, Development and Optimization of an Intelligent Tug Robot with an Automated Self-Aligning Locking Mechanism and Vision-Based Navigation

Autonomous Guided Vehicles (AGVs) play a crucial role in modern industrial logistics, where accurate navigation and reliable docking directly influence material-handling efficiency. However, conventional tug robots often experience docking errors due to mechanical misalignment and non-adaptive coupling mechanisms. This work presents the design, development, and optimization of an intelligent tug robot integrating vision-based navigation with an automated self-aligning locking mechanism. A ceiling-mounted camera combined with a convolutional neural network (CNN) is used for real-time detection and pose estimation of both the robot and the target cart. The navigation module computes the relative orientation and distance through geometric feature extraction, enabling high-precision approach and alignment. A conical male–female locking interface is developed to tolerate angular and translational misalignments, and its structural performance is validated using finite element analysis. Experimental results show that the robot achieves successful docking under ±۳۳° yaw deviation, ±۷۵° roll misalignment, and up to ۱۰ mm lateral offset. Integrating global machine vision with an optimized mechanical coupling significantly reduces docking time, improves repeatability, and enhances overall operational robustness. These findings demonstrate that the proposed tug robot provides a practical and cost-effective solution for autonomous load-handling applications in industrial environments.

Keywords:

AGV , Tug Robot , Vision-Based Navigation , CNN , Automated Locking Mechanism