On-Line Learning for Planning and Control of Underactuated Robots With Uncertain Dynamics

We present an iterative approach for planning and controlling motions of underactuated robots with uncertain dynamics. At its core, there is a learning process which estimates the perturbations induced by model uncertainty on active passive degrees freedom. The generic iteration algorithm makes use learned data in both phase, based optimization, control where partial feedback linearization dofs performed updated on-line. performance proposed shown comparative simulations experiments Pendubot executing various types swing-up maneuvers. Very few iterations are typically needed to generate dynamically feasible trajectories tracking that guarantees their accurate execution, even presence large uncertainties.