Towards low-complexity measurement-based feedback control
The control of quantum systems on the basis of continuous weak measurement has been developed along two main lines: so-called Markovian feedback, where the measurement output is just amplified and sent back as a drive to the system; and state-based feedback, where a full density operator is estimated with a quantum filter and well-behaved or optimal corrections are computed accordingly. In terms of controller complexity, these are two extreme cases. Markovian feedback has limited performance, a full filter can be computationally heavy even for moderate systems at ultra-fast timescales. We will report on our recent efforts to design controllers of intermediate complexity. In a first development, we introduce a low-dimensional dynamics to improve the performance of Markovian feedback. In a second development, we discuss a novel nonlinear structure in the quantum stochastic differential equation, which can enable to significantly reduce the dimension of the quantum filter.