quantum coherent feedback - qcc 2015qcc2015.weebly.com/uploads/6/6/5/5/6655648/gough.pdfcontrol and...
TRANSCRIPT
QUANTUM COHERENT FEEDBACK
John Gough
Quantum Structures, Information & Control, AberystwythQUANTUM CYBERNETICS & CONTROL
Nottingham University, January 2015
John Gough QUANTUM COHERENT FEEDBACK
What is Cybernetics? A transdisciplinary approach for exploring regulatory systems, their
structures, constraints, and possibilities. Cybernetics is relevant to the study of systems, such as mechanical,
physical, biological, cognitive, and social systems. Cybernetics is applicable when a system being analyzed is
involved in a closed signaling loop; that is, where action by the system generates some change in its environment
and that change is reflected in that system in some manner (feedback) that triggers a system change, originally
referred to as a ”circular causal” relationship. (Wikipedia)
Figure : Norbert Wiener (1948): cybernetics is ”the scientific study ofcontrol and communication in the animal and the machine”.
Root: Greek κυβερν meaning Helmsman, see also governor!
John Gough QUANTUM COHERENT FEEDBACK
What is Feedback?
If you have to ask, you’ll never know!
John Gough QUANTUM COHERENT FEEDBACK
What is Feedback? If you have to ask, you’ll never know!
John Gough QUANTUM COHERENT FEEDBACK
What is Feedback? If you have to ask, you’ll never know!
John Gough QUANTUM COHERENT FEEDBACK
Quantum Measurement-based Feedback Control
Partial information about the state of high Q cavity modes isobtained by measuring Rydberg atoms that are passed one-by-onethrough the cavity. The measurement results can be used to applya feedback action on the cavity mode.
I. Dotsenko, M. Mirrahimi, M. Brune, S. Haroche, J.-M. Raimond, and P. Rouchon, Quantum feedback by discrete
quantum nondemolition measurements: Towards on-demand generation of photon-number states Phys. Rev. A 80,
013805 (2009)
John Gough QUANTUM COHERENT FEEDBACK
(Autonomous) Quantum Coherent Feedback Control
John Gough QUANTUM COHERENT FEEDBACK
Why networks?
Processing of (classical) information on chip using photonics, YuriVlasov, CLEO 2012.
On chip quantum information processing: Old and New quantumteleportation devices (A. Furusawa)
John Gough QUANTUM COHERENT FEEDBACK
Connections through direct coupling
Given an system with Hamiltonian HS on Hilbert space hS , couplethe system directly to a second system (the governor) with Hilbertspace hG .
System Governor
The total evolution on hS ⊗ hG is of the form
H = HS ⊗ 1G + 1S ⊗ HG + V .
Design problems of this type first promoted by Seth Lloyd.
John Gough QUANTUM COHERENT FEEDBACK
Connections mediated by quantum fields
We begin with quantum Markov models. Each component isrepresented a system (S , L,H) as a single component with inputand output field:
System Hamiltonian H.
Coupling operator L between the system and the field.
Scattering operator S , unitary.
Figure : Gardiner’s input formalism Figure : (S,L,H)
John Gough QUANTUM COHERENT FEEDBACK
We consider formal “white noise” processes
[b (t) , b† (s)] = δ (t − s)
with
B (t) =
∫ t
0b (s) ds, B† (t) =
∫ t
0b† (s) ds.
It is possible to build a non-commutative version of the Ito calculus(Hudson and Parthasarathy, 1984; Gardiner and Collett,1985) on the Fock space over L2[0,∞) with respect todifferentials dB (t) and dB† (t), and we have
dB (t) dB† (t) = dt.
John Gough QUANTUM COHERENT FEEDBACK
A unitary system + noise dynamics:
dU =
{L⊗ dB† − L† ⊗ dB −
(1
2L†L + iH
)⊗ dt
}U
future pointing differentials!
The flow of system observables jt (X ) = U† (t) [X ⊗ 1] U (t):
Heisenberg-Langevin equations of motion
djt (X ) = jt (LX ) dt + jt([X , L])dB† + jt([L†,X
])dB,
where LX = 12 [L†,X ]L + 1
2 L† [X , L]− i [X ,H].
Implies the master equation for %t : 〈jt(X )〉 = tr{%tX}.John Gough QUANTUM COHERENT FEEDBACK
For example, we may have an optical cavity, H = ~ωa†a, withcoupling
L =√γa.
Figure : Absorption of field quanta Figure : emission of field quanta
Leads to
d%tdt
= γ
(a%ta
† − 1
2a†a%t −
1
2%ta†a
)+ iω[%t , a
†a].
John Gough QUANTUM COHERENT FEEDBACK
Linear Quantum Systems (Yanagisawa and Kimura, 2003)
John Gough QUANTUM COHERENT FEEDBACK
Handling quantum signals coherently!
Beamsplitters: The S in the ”SLH” ...[Bout
1
Bout2
]=
[S11 S12
S21 S22
] [B1
B2
].
John Gough QUANTUM COHERENT FEEDBACK
Cascades: Systems in Series
Output fields Bout (t) = U†t (1⊗ B (t))Ut
dBout (t) = jt(S)dB (t) + jt(L)dt.
We generalize the notion of cascade introduced by H.J. Carmichael†.
dB(2)out = S2dB
(2)in + L2dt
= S2(S1dB(1)in + L1dt) + L2dt
= S2S1B(1)in + (S2L1 + L2)dt
† H.J. Carmichael, Phys. Rev. Lett., 70(15):2273 2276, 1993.
John Gough QUANTUM COHERENT FEEDBACK
The Series Product
The Series Product† gives the rule for nonlinear cascaded quantumMarkov systems in the instantaneous feedforward limit isequivalent to the single component:
(S2, L2,H2) C (S1, L1,H1) =(S2S1, L2 + S2L1,H1 + H2 + Im
{L†2S2L1
}).
John Gough QUANTUM COHERENT FEEDBACK
Bilinear Control Hamiltonian∗
Based on H. M. Wiseman and G. J.Milburn. All-optical versuselectro-optical quantum-limitedfeedback.Phys. Rev. A, 49(5):41104125, 1994.
(I , u (t) , 0) C (−I , 0, 0) C (I , L, 0) C (−I , 0, 0) C
(I ,−u (t) , 0) C (I , L, 0) = (I , 0,H (t))
where
H (t) = Im{L†u(t)} =1
2iL†u (t)− 1
2iLu (t)∗ .
∗ J. G., Construction of bilinear control Hamiltonians using the seriesproduct and quantum feedback Phys. Rev. A 78, 052311 (2008)
John Gough QUANTUM COHERENT FEEDBACK
Components in-loop
Back to the model considered by Yanagisawa & Kimura:Example
beamsplitter S =
[S11 S12
S21 S22
],
and in-loop component (S0, L0, 0):
dB2 = S0dBout2 + L0dt = S0(S21dB1 + S22dB2) + L0dt
⇒ dBout1 = S11dB1 + S12dB2 ≡ S0dB1 + L0dt
where
S0 = S11 + S12(I − S0S22)−1S0S21, L0 = S12(I − S22)−1S0L0.
Equivalent component (S0, L0, H0):
John Gough QUANTUM COHERENT FEEDBACK
Networks Rules
More generally how do we buildarbitrary networks from multiplecomponents.
Rule # 1: Open loop system in parallel
�nj=1 (Sj , Lj ,Hj) =
S1 0 0
0. . . 0
0 0 Sn
, L1
...Ln
,H1 + · · ·+ Hn
.
John Gough QUANTUM COHERENT FEEDBACK
Feedback Reduction Formula:
Rule # 2: Open loop system in parallel Feedback Reduction
B =
[Be
Bi
]L =
[Le
Li
]
S =
[See Sei
Sie Sii
]X
external external
internal
S fb = See + Sei
(X−1 − Sii
)−1Sie ,
Lfb = Le + Sei
(X−1 − Sii
)−1Li ,
H fb = H + ImL†eSei
(X−1 − Sii
)−1Li
+ ImL†i Sii
(X−1 − Sii
)−1Li .
John Gough QUANTUM COHERENT FEEDBACK
QHDL: a hardware description language for QFNs
Computer-aided schematic capture workflow for modelling andsimulating multi-component photonic circuit
Figure : Pseudo-NAND circuit schematic (a) as created with scheme andits device symbol embedded as a component in an SR-NAND-latchcircuit (b). John Gough QUANTUM COHERENT FEEDBACK
Quantum filtering
Non-demolition SchemeConditional expectation ontothe measurement algebra Jtgenerated by J(s), 0 ≤ s ≤ t.
πt(X ) = E[U∗t (X ⊗ I )Ut | Jt ]
(Quadrature Measurement)
dπt(X ) = πt(LX )dt + {πt(XL + L∗X )− πt(L + L∗)πt(X )}×[dJ(t)− πt(L + L∗)dt].
(Photon Counting Measurement)
dπt(X ) = πt(LX )dt +
{πt(L∗XL)
πt(L∗L)− πt(X )
}×[dJ(t)− πt(L∗L)dt].
John Gough QUANTUM COHERENT FEEDBACK
The Separation Principle
To minimize an expectation cost based on a fixed measurementprocedure, one has a Hamiltonian-Jacobi-Bellman principle1.There is a quantum analog of the separation principle for controllerdesign2.
1 J. G., V.P. Belavkin, O.G. Smolyanov, Hamilton-Jacobi-Bellman equations for Quantum Filtering and Control, J.Opt. B: Quantum Semiclass. Opt. 7 S237-S244 (2005)
2 L. Bouten, R van Handel, Quantum Stochastics and Information: Statistics, Filtering and Control (V. P. Belavkin
and M. I. Guta, eds.), World Scientific, (2008)
John Gough QUANTUM COHERENT FEEDBACK
Risk Sensitive Measurement-Based Feedback Control
The design of the optimal controller can be split into
the estimation stage (calculating the least squares estimatefor the state, i.e. the conditional state %t),
the actuation stage (determining the control policy based on%t).
However a LEQG (linear exponential quadratic gaussian) versionalso exists in the quantum case3.
The optimal estimate %t depends on the past measurement recordand the control objective!
3 M.R. James, Risk-Sensitive Optimal Control of Quantum Systems, Physical Review A, 69, 032108 (2004)
M.R. James, A Quantum Langevin Formulation of Risk-Sensitive Optimal Control, J. Opt. B: Quantum Semiclass.
Opt. 7 S198–S207 (2005)
John Gough QUANTUM COHERENT FEEDBACK
Non-Vacuum input
J.G., Guofeng Zhang, Generating Nonclassical Quantum InputField States with Modulating Filters, arXiv:1404.3866
John Gough QUANTUM COHERENT FEEDBACK
Quantum Memories:
H.I. Nurdin, J.G., Modular Quantum Memories Using PassiveLinear Optics and Coherent Feedback, arXiv:1409.7473, to appearQuantum Info. Processing.
John Gough QUANTUM COHERENT FEEDBACK
Quantum Coherent Control in Quantum Transport
John Gough QUANTUM COHERENT FEEDBACK
Quantum Feedback Networks in Quantum Transport
John Gough QUANTUM COHERENT FEEDBACK