Quantum filtering using POVM measurements

The objective of this work is to develop a recursive, discrete time quantum filtering equation for a system that interacts with a probe, on which measurements are performed according to the Positive Operator Valued Measures (POVMs) framework. POVMs are the most general measurements one can make on a quantum system and although in principle they can be reformulated as projective measurements on larger spaces, for which filtering results exist, a direct treatment of POVMs is more natural and can simplify the filter computations for some applications. Hence we formalize the notion of strongly commuting (Davies) instruments which allows one to develop joint measurement statistics for POVM type measurements. This allows us to prove the existence of conditional POVMs, which is essential for the development of a filtering equation. We demonstrate that under generally satisfied assumptions, knowing the observed probe POVM operator is sufficient to uniquely specify the quantum filtering evolution for the system.

💡 Research Summary

This paper develops a foundational framework for discrete-time quantum filtering when measurements on a probe are described by Positive Operator-Valued Measures (POVMs), the most general class of quantum measurements.

Traditional quantum filtering theory, crucial for state estimation and feedback control, has largely been built upon the assumption of projective measurements (PVMs). While any POVM can be mathematically reformulated as a PVM on a larger Hilbert space, this lift is non-unique and can be conceptually cumbersome. This work directly addresses POVMs, offering a more natural and potentially simplified formalism for applications where measurements inherently correspond to POVMs, such as approximate position or phase measurements.

The core challenge in moving to POVMs lies in their non-unique “instruments.” A POVM defines outcome probabilities but does not uniquely specify the post-measurement state; for that, an associated instrument (a completely positive map) is needed, and infinitely many instruments can correspond to the same POVM. The paper’s key theoretical innovation is formalizing the notion of “strongly commuting instruments.” Two instruments strongly commute if the operator sequences representing them commute pairwise. The authors prove that under this condition, one can meaningfully define joint measurement statistics for their corresponding POVMs and, consequently, the existence of conditional POVMs.

The filtering application considers a system interacting with a probe. After interaction, a POVM measurement is performed on the probe. The paper demonstrates that any physically reasonable instrument acting only on the probe strongly commutes with any instrument acting only on the system. This fundamental commutativity is the linchpin that allows the construction of a recursive filtering equation. It enables the calculation of the conditional expectation of the system state given the observed probe POVM outcome.

A major practical result is that the derived filtering equation depends solely on the observed POVM operator on the probe. It does not require knowledge of the specific instrument implementing that POVM or other POVM elements. This is significant because experimentally identifying the full instrument (which requires post-measurement state tomography) is often infeasible, whereas the POVM can be deduced from measurement statistics alone. Therefore, the theory provides a self-contained and experimentally accessible filtering recipe directly at the POVM level.

In summary, this work generalizes quantum filtering theory to the broadest measurement class by leveraging the concept of strongly commuting instruments. It bypasses the non-unique and sometimes problematic lift to PVMs, delivering a concise filtering equation that uses only POVM data, thereby enhancing both theoretical elegance and practical applicability in quantum feedback control and real-time estimation tasks.