Learning a Kinodynamic Trajectory Manifold for Impact-Aware Compliant Catching of Fast-Moving Objects
For robotic manipulation, this provides a real-time compliant catching method for fast-moving objects, though the approach is incremental.
The paper tackles fast catching of free-flying objects under kinodynamic constraints and impact uncertainty. Using RL in simulation, they learn a low-dimensional trajectory manifold that maps object state to a reference trajectory, achieving successful catches without online optimization.
Fast catching of free-flying objects is difficult because of short reaction time, impact uncertainty, and kinodynamic constraints. We use reinforcement learning in simulation to collect successful catching trajectories and learn a low-dimensional kinodynamic trajectory manifold. At run time, the estimated object initial state is mapped directly to a reference catching trajectory without online nonlinear optimization. The trajectory is tracked with compliant control near contact for improved impact absorption and capture stability.