Abstract
Finally, the sixth and final lesson dealt with quantum feedback. This process generalizes to a quantum-coherent system the process of controlling
of a classical system by feedback. The need for this type of control can be understood from the elementary example of maintaining the temperature in a room, or stabilizing an aircraft at a given altitude (autopilot). Spontaneously, the variable to be controlled is not subject to any restoring force or damping, and can therefore drift over time. The aim of control is to prevent this drift. The feedback loop, which introduces additional degrees of freedom coupled to the controlled system, constitutes an artificial restoring force and damping. There are in fact two types of feedback control. The first, used for example in ion trapping, consists in subjecting the system to stabilizing oscillating forces. But in this autonomous control scheme, there is no measuring device as such, and its possibilities are limited. The second type of control involves the active participation of a measuring device. This records the evolution of the controlled variable over time. The result is then sent to a computing device which estimates the state of the system and derives, from the value of a setpoint, the value of the corrective force to be applied to keep the system on course. The problem of control becomes subtle when applied to a quantum-coherent system that is perturbed by measurement. For a quantum system, the uncertainty principle invalidates the assumption that, in a classical system, the noise added by the measurement can be completely independent of the noise causing the system to deviate from its trajectory.
