Continuous Dynamical Decoupling with Bounded Controls

Abstract

We develop a theory of continuous decoupling with bounded controls from a geometric perspective. Continuous decoupling with bounded controls can accomplish the same decoupling effect as the bang-bang control while using realistic control resources and it is robust against systematic implementation errors. We show that the decoupling condition within this framework is equivalent to average out error vectors whose trajectories are determined by the control Hamiltonian. The decoupling pulses can be intuitively designed once the structure function of the corresponding SU(n) is known and is represented from the geometric perspective. Several examples are given to illustrate the basic idea. From the physical implementation point of view we argue that the efficiency of the decoupling is determined not by the order of the decoupling group but by the minimal time required to finish a decoupling cycle.