In classical engineering, feedback control is straightforward: a sensor measures the state of a system, a classical processor computes an error signal, and an actuator applies a corrective force. However, when attempting to control quantum systems, this paradigm encounters a fundamental physical barrier.
In the quantum domain, measurement is inherently destructive. Observing a quantum state causes wavefunction collapse, destroying delicate superpositions and introducing stochastic projection noise. Furthermore, converting a quantum signal into classical data and back again introduces significant time delays and latency. As noted by Daoyi Dong et al. in their roadmap paper, Quantum estimation, control and learning: opportunities and challenges, bypassing this classical measurement bottleneck is essential for exceeding standard quantum performance limits.
Coherent feedback control offers an elegant solution to this dilemma. Instead of using a classical observer, a coherent feedback loop utilizes an auxiliary quantum system as the controller. The target system (the "plant") and the controller interact directly, allowing information and energy to flow continuously between them entirely within the quantum domain. Because no classical measurement occurs, the system avoids wavefunction collapse, minimizes latency, and preserves quantum coherence throughout the entire control loop.
The Architecture of a Fully Quantum Loop
To understand why coherent feedback works, we must treat quantum information not as abstract data, but as a physical resource. In a traditional measurement-based setup, quantum information is extracted, collapsed into classical bits, and processed. In a coherent feedback loop, the quantum information remains unmeasured and is routed directly into the controller.
As outlined in a comprehensive review by Lincheng Wang et al., Closed‐Loop and Robust Control of Quantum Systems, this interaction typically occurs via direct Hamiltonian coupling or through traveling fields, such as optical signals. For example, in quantum optics, a system's output field can be routed through optical components—such as beam splitters, phase shifters, and optical cavities—and fed directly back into the system.
By manipulating these optical paths, researchers can achieve control performance that surpasses classical limits. Guofeng Zhang et al. demonstrated this in Coherent Feedback Control of Linear Quantum Optical Systems via Squeezing and Phase Shift, showing that incorporating quantum optical manipulations like "squeezing" directly into the feedback loop allows the system to perform better than what is theoretically possible using standard vacuum states.
From an information-theoretic perspective, keeping the loop fully quantum enables deterministic, nondestructive control actions. Daoyi Dong et al. explored this in INFORMATION-TECHNOLOGY APPROACH TO QUANTUM FEEDBACK CONTROL, illustrating how advanced quantum protocols like teleportation and cloning can be embedded directly within a feedback architecture to balance information gain against state disturbance.
Designing for Physical Realizability
In classical control theory, a controller can be any mathematical algorithm that stabilizes a set of differential equations. In coherent quantum control, however, the controller is a physical quantum system. This imposes a strict mathematical constraint known as physical realizability: the controller's equations of motion must correspond to a valid open quantum system that obeys the fundamental laws of quantum mechanics, such as preserving commutation relations.
To design controllers that are physically realizable by default, researchers often embed the control objectives directly into the system's joint Hamiltonian. As shown by Chengdi Xiang et al. in A Popov approach to performance analysis and coherent guaranteed cost control for uncertain quantum systems, augmenting the nominal system Hamiltonian with a dedicated control Hamiltonian component ensures that the resulting feedback mechanism satisfies these strict physical laws without requiring external measurement.
Furthermore, the internal dynamics of the quantum controller dictate its capabilities. In Static and dynamic coherent robust control for a class of uncertain quantum systems, Chengdi Xiang et al. compared static controllers (which merely route signals) to dynamic controllers (which possess their own internal quantum states). They proved that dynamic coherent controllers offer superior stability and performance when protecting a system against uncertain Hamiltonian perturbations, as their internal degrees of freedom allow for more sophisticated, time-dependent compensation.
Engineering Robustness Against Noise and Faults
Quantum systems are exceptionally fragile, meaning a primary goal of any control loop is to mitigate environmental noise and hardware fluctuations. Because coherent feedback loops lack a classical processor to filter out anomalies, robustness must be engineered directly into the quantum dynamics.
To achieve this, control theorists have adapted classical robust control frameworks—such as $H^\infty$ control, the small-gain theorem, and Popov methods—to the quantum domain. In Performance Analysis and Coherent Guaranteed Cost Control for Uncertain Quantum Systems Using Small Gain and Popov Methods, Chengdi Xiang et al. established a framework to design coherent controllers that guarantee stability and minimize a defined performance cost function, even when the underlying quantum Hamiltonian faces unknown quadratic or non-quadratic perturbations (see also Performance Analysis and Coherent Guaranteed Cost Control for Uncertain Quantum Systems).
+------------------+ Quantum Signal (e.g., Optical Field) +----------------------+
| | ---------------------------------------> | |
| Target Quantum | | Quantum Controller |
| System (Plant) | <--------------------------------------- | (Auxiliary System) |
| | Coherent Feedback Interaction | |
+------------------+ +----------------------+
^ ^
| |
+------------------- Joint Unitary Evolution -------------------+
(No Classical Measurement)
This robust framework is particularly powerful for handling discrete, unpredictable hardware faults. For instance, linear quantum systems can suffer from sudden structural changes known as Markovian jumping faults. Yanan Liu et al. developed a fault-tolerant coherent $H^\infty$ control strategy in Fault-Tolerant Coherent $H^\infty$ Control for Linear Quantum Systems, demonstrating that a fully quantum controller can attenuate external disturbances and maintain system integrity despite these jumps.
In practical optical setups, this approach allows for direct, real-time stabilization. In Fault-tolerant H∞ control for optical parametric oscillators with pumping fluctuations, Daoyi Dong et al. applied these principles to stabilize Optical Parametric Oscillators (OPOs). By processing quantum signals directly in the optical domain, their coherent feedback loop successfully mitigated amplitude and phase fluctuations in the driving pump laser without introducing classical measurement delays.
Applications Beyond Stabilization
While stabilization and noise rejection are the primary drivers of coherent feedback, the ability to manipulate quantum states without measurement opens up unique capabilities across quantum technologies.
Environmental Decoupling via Chaos
Coherent feedback can be used to intentionally alter a system's frequency response to shield it from its surroundings. In Noise suppression of on-chip mechanical resonators by chaotic coherent feedback, Nan Yang et al. connected an on-chip mechanical resonator to a secondary dynamical system that fed back a continuous, broadband chaotic signal. This measurement-free "kicking" broadened the resonator's frequency response, effectively decoupling it from its thermal environment and allowing it to serve as a high-fidelity quantum memory.
Universal Quantum Computation
The principles of coherent feedback also scale up to quantum computing architectures. In the model of Ancilla-Controlled Quantum Computation (ACQC) proposed by Erika Andersson et al. in Universal quantum computation by the unitary control of ancilla qubits and using a fixed ancilla-register interaction, an entire computational register is controlled through a single, fixed unitary interaction with an addressable ancilla qubit. Because the register is manipulated solely through its coherent feedback loop with the ancilla, universal quantum logic is achieved entirely without intermediate measurements or direct external access to the data qubits.
Key Foundational Work
INFORMATION-TECHNOLOGY APPROACH TO QUANTUM FEEDBACK CONTROL (2006) — contributed 2 concept elements to this direction.
Coherent Feedback Control of Linear Quantum Optical Systems via Squeezing and Phase Shift (2012) — contributed 1 concept elements to this direction.