The last frontiers of robotic applications present increasingly restrictive constraints of control robustness and performance. Concerning this critical requirement, we provide a simple approach leveraging on a switching control policy that we propose optimizing a quadratic performance index that relates to the \(\mathcal {L}_2\) norm of the tracking error. The approach allows the definition of an optimal switching strategy orchestrating the agent controller, bumpless transfer and reference governor filters, combined with state resets enhancing the performances of the resulting closed-loop hybrid system, which is guaranteed to be globally asymptotically stable while robustness is guaranteed by Lyapunov arguments. We also provide the dual formulation applied to the estimation problem.