A Swing-Aware Local Planning and MPC Tracking Method for Overhead Cranes

Abstract

This brief studies obstacle-avoiding reference generation and swing-suppressed tracking for overhead cranes under small-angle and constant-rope-length assumptions. In a cluttered workspace, a collision-free trolley path may still be unsuitable if its timing and acceleration profile excite swing or leave insufficient room for braking. To address this issue, a timed elastic band (TEB)-style local planner is formulated over position, velocity, acceleration, and transition-time variables. For each candidate local trajectory, the acceleration sequence is propagated through a linearized swing model, and the predicted swing is penalized in the local objective. The obstacle influence radius is also updated along the same candidate trajectory according to a learned braking-distance estimate, which depends on the local speed and predicted swing state. The optimized time-parameterized reference is then tracked by two constrained axis-wise MPC controllers. The static-obstacle simulations, one moving-obstacle case, and laboratory-scale experiments show that the proposed scheme reduces payload swing and avoids overly conservative detours within the tested operating regime.