Relative‑motion dynamics and autonomous docking control for a levitated shuttle–robot interface

Abstract

Autonomous docking between a magnetically levitated shuttle and a robotic end-effector presents unique safety and precision challenges that existing research does not adequately address. While docking methodologies are well established in spacecraft rendezvous and docking, no structured framework exists for hybrid systems in which one platform is freely levitated and the other is a ground-fixed manipulator. This work addresses that gap by adapting spacecraft rendezvous principles to a terrestrial planar robotic setting.

Three contributions are presented. First, a relative-motion docking model is developed that defines conical approach corridors, spherical keep-out zones, and phase-dependent velocity constraints to ensure geometrically safe trajectories. Second, a finite state machine is implemented in a Programmable Logic Controller using Structured Text and Python, governing the full docking sequence through sequential phases, deterministic transition conditions, and abort logic. Third, the proposed method is experimentally validated on an ACOPOS 6D magnetic levitation platform and compared against a simple straight-line baseline across ninety trials at three distinct starting positions.

Both controllers achieved a 100% docking success rate. Final position errors were approximately 1.33×10^-7 m, and no corridor or keep-out violations occurred during nominal trials. Docking times were comparable between the methods, with differences attributable primarily to intentional dwell periods in the optimized controller. The results demonstrate that spacecraft-inspired corridor geometry and phase-structured execution logic can be implemented reliably on a planar levitated system without degrading performance, and suggest strong potential for scaling these methods to higher degree-of-freedom or safety-critical robotic docking tasks.