Industrial autonomous robot tractors are increasingly deployed in logistics yards for trailer transport and docking. Yet, most research and systems prioritize safety and automation performance, while overlooking the human aspects involved in supervising these tractors. This imbalance risks reducing efficiency and safety in practice. We argue that supervision of autonomous tractors must be recentered on the human aspects of logistics through Human-Computer Interaction, ergonomics, and Human-in-the-loop principles. Drawing on an activity analysis of workflows in a representative logistics hub, we identify critical tasks and propose three design considerations for a human-centered design of supervision systems of autonomous robot tractors. We outline a lightweight conceptual framework that embeds transparency, explainability, ergonomics, and trust calibration into supervisory interfaces. We believe that the autonomous robot tractors can only be safely and effectively integrated into logistics operations by reframing supervision as a human-centered issue.