Caster aware control improves robot motion in tight turns
MSC: Multi-Stage Caster-aware Control Framework with Uncertainty-Aware NMPC and RL-based Execution Layer
RoboticsMachine Learning
Summary
Swiveling casters on robots make turning and carrying loads easier, but sometimes the wheels don’t line up right, causing jerky or inaccurate movement. The authors developed a new control system called MSC that acts in two stages to handle this problem better: first by predicting and adjusting for uncertain wheel positions when planning moves, and second by using a learning method to convert commands into wheel speeds that fit actual contact conditions. They tested this system both in simulation and on real robots, showing it keeps movements smoother and more accurate during tricky maneuvers.
What this means in practice
- •For industrial robot developers: Enhance autonomous mobile robots to maintain precise movement and load handling during complex, low-speed turns with improved caster wheel control.$Commercial implications: Allows manufacturers to offer robotic platforms with smoother, more reliable navigation in cluttered environments where caster misalignment reduces performance.
- •For autonomous vehicle engineers: Improve control algorithms for slow-speed maneuvers in self-driving delivery vehicles to reduce slippage and tracking errors caused by free-swiveling caster wheels.
Authors
Jinho Park, Jeongtae Lee, Jeonghyeok Lim, Keun Ha Choi, Kyung-Soo Kim
Abstract
Freely swiveling casters are widely adopted in industrial AMRs/SDVs for their mechanical simplicity and load capacity. However, caster misalignment during low-speed or high-curvature maneuvers induces strong contact nonlinearities, widening the planning-execution gap between commanded $(v, ω)$ and realized motion. Existing caster-aware strategies are highly sensitive to state-estimation reliability, causing unstable execution in uncertain transition regimes. To address this, we present Multi-Stage Caster-aware Control (MSC), a planning-execution integrated hierarchical framework featuring two core mechanisms. First, an uncertainty-aware NMPC (CU-MPC) dynamically schedules caster-aware costs and input regularization using predictive uncertainty to suppress overly aggressive maneuvers. Second, a reinforcement learning-based execution layer estimates and predicts caster states under partial observability, transforming high-level velocity commands into physically executable wheel speeds adapted to current contact conditions. Validations in Isaac Sim and real robots demonstrate that MSC offers a clear comparative advantage over existing decoupled compensation approaches by significantly improving caster-state estimation accuracy, minimizing misalignment, and enhancing tracking robustness in caster-sensitive maneuvers.