Multi-Layer Vertex Refinement Algorithm for Online Palletization of Boxes With Varying Dimensions

This paper presents an online palletization algorithm, designed primarily for robotic applications, designed to handle varying-sized boxes in industrial environments. The proposed algorithm leverages the capabilities of articulated robots and 3D camera technology to perform efficient and adaptive palletization. It addresses the significant challenge of accommodating diverse box dimensions through a dynamic, real-time algorithm, that utilizes data from a 3D camera to perceive the size and shape of each incoming box accurately. Subsequently, the robot employs a carefully optimized manipulation strategy to grasp and place boxes on the pallet in an optimal arrangement. The algorithm’s adaptability is emphasized as it continuously refines its palletizing strategy in response to the changing characteristics of incoming boxes. Experimental results emphasize the algorithm’s efficacy, demonstrating an average pallet space utilization between 75.63% and 82.88% for common pallet sizes, with a peak efficiency reaching 92.77%. This research contributes to the evolving field of robotics and automation, offering a versatile solution for palletizing tasks where the box dimensions can vary significantly.