Optical Coherent Detection
Optical Coherent Detection
Optical Coherent Detection
K. M. Mehta
Abstract: Coherent detection of optical signal has once again gained the attention of research
community because of its high sensitivity and ability to compensate linear transmission impairments.
Researchers have demonstrated polarization multiplexed transmission system compatible with 100-
Gb/s Ethernet standard with these receiver. This article discusses about basics of coherent reception
and digital signal processing involved.We also various types of receiver architectures and recent
demonstrations/experiments.
1. Introduction
A detection system in which phase coherence of the optical carrier plays an important role is
called coherent detection system. These systems are well known in the context of radio and
microwave communication but until recent its use in commercial systems has been hindered by
the additional complexity, due to the need to track the phase and the polarization of the incoming
signal. For optical communication, intensity modulated direct detection(IM/DD) systems remained
dominant in term of commercial deployment. The drive for higher performance and high data rate
in optical fiber systems has renewed interest in coherent detection.
Coherent detection has following advantages over direct detection[1][2]:
• Receiver is sensitive to the phase as well as the amplitude of the optical wave, therefore
linear impairments,including chromatic dispersion and polarization-mode dispersion, can be
compensated quasi-exactly using electronic filters.
• Higher data rate of the order of 111 Gbps can be achieved exploiting all parameters (inten-
sity,angle,polarization) of modulation.
• Coherent detection offers an inherent ultra narrow optical filtering capability useful for dense
wavelength-division multiplexing
• We can have up to 20 dB improvement in receiver sensitivity when compared with IM/DD
systems.
• Some nonlinear impairments, such as intra-channel four-wave mixing and nonlinear phase
noise, can be compensated partially.
Importance in case of long-haul communication systems: For a long-haul optical fiber system
it is required to transmit the highest data throughput over the longest distance without signal
regeneration. Given constraints on the bandwidth imposed by optical amplifiers, it is important to
maximize spectral efficiency. Also given constraints on signal power limited by fiber nonlinearity,
we need a power efficient system which minimizes the required average transmitted energy per
bit. Coherent detection with polarization multiplexing achieves both goals as in-phase (I) and
quadrature-phase (Q) signals in the two orthogonal field polarizations allows information to be
encoded in all the available degrees of freedom.
Importance of Digital signal processing: In this paper we will discuss various digital signal
processing techniques. A DSP-based receiver is highly advantageous because adaptive algo-
rithms can be used to compensate time-varying transmission impairments. Advanced forward
error-correction coding can also be implemented. Moreover, digitized signals can be delayed,
split and amplified without degradation in signal quality[3].
This paper is organized as follows: In Section 2, we review basics of coherent detection and
various receiver architectures . In Section 3, we discuss about digital signal processing algorithm
for carrier phase recovery, equalization and detection. In Section 4, we discuss various demon-
strations and research experiments and finally we conclude the discussion in Section 5.
Since a photo detector responds to the optical intensity, the optical power incident at the photo
detector is given by:
2
P = K |Es + ELO | (3)
where
and
ωIF is also known as intermediate frequency.There are following two types of coherent receiver
based on value of ωIF :
• Homodyne receiver.
• Heterodyne receiver.
Typically, Ps PLO and consider the case in which the local-oscillator phase is locked to the
signal phase so thatφs − φLO . The homodyne signal is then given by
p
I(t) = 2R Ps PLO (8)
The main advantage of homodyne detection can be seen from Eq.(8) If we note that the signal
current in the direct-detection case is given by Idd(t) = RPs (t). Denoting the average optical
power by P¯s , the average electrical power is increased by a factor of 4PLO /Ps with the use of
homodyne detection. Since PLO can be made much larger than Ps , the power enhancement can
exceed 20 dB. Although in this case shot noise is also enhanced, homodyne detection improves
the signal-to-noise ratio (SNR) by a large factor.
A disadvantage of homodyne detection is its phase sensitivity.In practice, both φs and φLO
fluctuate with time in a random manner. However, their difference φs − φLO can be forced to
remain nearly constant through an optical phase-locked loop. In addition,exact matching of the
transmitter and local-oscillator frequencies is required.
Since PLO Ps in practice, the direct-current (DC) term is nearly constant and can be removed
easily using bandpass filters. The heterodyne signal is then given by the following alternating-
current (ac) term:
p
Iac (t) = 2R Ps PLO cos (ωIF t + φs − φLO ) (10)
Similar to the case of homodyne detection, the local oscillator amplifies the received signal by a
large factor, thereby improving the SNR. However, the SNR improvement is lower by a factor of 2
(or by 3 dB) compared with the homodyne case. This reduction is called as “heterodyne detection
penalty”. Because of the ac nature of Iac , the average signal power is reduced by a factor of 2
2
when Iac is averaged over a full cycle at the intermediate frequency.
The advantage gained at the expense of the 3-dB penalty is that the receiver design is consid-
erably simplified and the line width requirements are quite moderate. These features makes the
heterodyne detection scheme quite suitable for practical implementation.
RP̄s
SN R ≈ (14)
q∆f
The use of coherent detection allows one to achieve the ‘‘shot-noise limit” even for PIN receivers
whose performance is generally limited by thermal noise. Moreover, in the case of avalanche photo
diode (APD) receivers, this limit is realized without adding any excess shot noise.
2 2
P1b (t) = |As (t)| + |ALO (t)| − 2Re{As (t)A?LO (t)exp (iωIF t + iφs − iφLO )} (16)
and the balanced photo detector output is, therefore, proportional to
∆P1 (t) = P1a (t) − P1b (t) ≈ 4Re{As (t)A?LO (t)exp (iωIF t + iφs − iφLO )} (17)
The DC term is eliminated completely during the subtraction process when the two branches
are balanced (each branch receives equal signal and local-oscillator powers). More importantly,
the intensity noise associated with the DC term is also eliminated during the subtraction process
since intensity fluctuations in the two branches are perfectly correlated and cancel out. The quarter
wave shift in LO path of the lower arm of the 90◦ hybrid means that the balanced photo detector
in the lower arm responds to
3.2. Equalization
Electronic equalization is used to compensate for linear transmission impairments and recover
the polarization multiplexed data. This usually involves a filtering operation[4]. Equalization can
be partitioned into two distinct stages.
• Chromatic Dispersion compensation
• Polarization dependent impairment (Polarization rotation and PMD)compensation
−jDλ2 2
A(z, ω) = A(0, ω)exp ω z (22)
4πc
This can be viewed as LTI system with frequency response of
−jDλ2 2
H(z, ω) = exp ω z (23)
4πc
and thus frequency response of dispersion compensating filter is given by Hc (z, ω) = 1/H(z, ω).Such
a filter can be implemented using FIR or IIR system(as in case of [6]). Let us discuss design of
FIR filter. Time domain impulse response of Hc (z, ω) is given by:
r
−jπct2
jc
hc (z, t) = exp (24)
Dλ2 z Dλ2 z
It can be seen that this impulse response is of infinite duration, non-causal and since it passes all
frequencies, for a finite sampling frequency aliasing will occur.By truncating this to finite length by
using a Kaizer window we can solve these problems. Length of the window(which is proportional
to N taps of an FIR filter) can be calculated using following criterion as specified in [5]
Dλ2 z Dλ2 z
− ≤t≤ (25)
2cTs 2cTs
where Ts is sampling interval.Filter coefficients of N tap FIR filter are given by
r
jcTs2 πcTs2 2
ak = exp −j k (26)
Dλ2 z Dλ2 z
where
Dλ2 z
N N
− ≤k≤ and N = 2 +1 (27)
2 2 2cTs2
This filter can be implemented in FPGA using shift registers or with FFT algorithm.
and
M
X −1
0 T T
y (k) = Hyx Xp + Hyy Yp = hyx (m)xp (k − m) + hyy (m)yp (k − m) (29)
m=0
where hxx ,hxy ,hyx ,hyy are all adaptive filters of length M .Various methods have been suggested
for calculating coefficients of these filters but we will look into ‘‘Constant Modulus Algorithm”(CMA)
suggested by Dominique[8].This algorithm exploits the constant modulus property of QPSK mod-
ulated signal. Let us assume unit amplitude of the modulated signal.The equalizer will try to
minimize mean square error x and y given by:
2 2
x = 1 − |x0 | and y = 1 − |y 0 | (30)
giving following four criterion:
d 2y d 2y
d 2x d 2x
=0, =0, =0, =0 (31)
dHxx dHxy dHyx dHyy
To determine optimum tap weights stochastic gradient algorithm is used.In the initialization stage
all tap weights are set to zero with the exception of central tap of hxx and hyy . Once the equalizer
has converged it moves to decision directed mode using DD-LMS algorithm. The filter coefficients
are now adapted as:
2
µ d |x |
Hxx → Hxx − = Hxx + µx x¯p (32)
2 dHxx
where x¯p is complex conjugate of xp and µ is a convergence parameter. and x = dx − x0 with
dx being the symbol closest to x0 . Similarly filter coefficients of Hxy ,Hyx ,Hyy are calculated on
the fly.
estimate, as shown in the following. Coherent detection using such a method for carrier phase
estimation will offer significant improvement in BER performance over differential detection[11].
The equalized signal samples(x0 (n) = Equalized (xi + jxq )) and (y 0 (n) = Equalized (yi + jyq ))
in both the polarization are processed with the DSP circuit shown in Fig. 5.Let us call x0 (n) or
y 0 (n) by E(n) for explaining the algorithm. The processing module consists of a demultiplexing
unit, several processing units (PUs) that estimate the phase offset, and a remultiplexing unit. The
stream of E(n) are demultiplexed into M blocks of N complex samples (as shown in Fig. 5)
before being sent to the PUs.When ASE noise is the dominant noise N should be as large as
possible. The core processing functions residing in the PUs can demodulate the data in parallel
and independently from each other, which allows this architecture to support high symbol rates.
A schematic of a PU is shown in Fig. 6.E(i)(i = 1, ..., N ) to the fourth power cancels the
QPSK phase modulation θs (= 0, π/2, π, π/2), sinceE 4 ∝ exp[j(4θs )]. The complex amplitudes E 4
are summed, so that the phase is averaged over the entire block. The phase of the resulting
complex amplitude is divided by 4, leading to a phase correction for the k th block θ0 (k) that lies
between π/4 and π/4. The phase correction θ0 (k) is, thus, given as
"N #
0 1 X
4
θ (k) = arg E (kN + i) (33)
4 i=1
This phase error can be subtracted form each symbol of the block to get modulated symbol
back.It may be noted that differential phase encoding solves the problem of phase ambiguity as
in this case absolute phase is not relevant and the information is encoded in phase difference of
two consecutive symbols.
References
[1] G. P. Agrawal, Fiber-Optic Communications Systems, Third Edition. John Wiley and Sons, 2002.
[2] D. J. F. B. J. M. K. Ezra Ip, Alan Pak Tao Lau, “Coherent detection in optical fiber systems,” Optical express, vol. 16,
no. 2, 2008.
[3] M. G. Taylor, “Coherent detection method using dsp for demodulation of signal and subsequent equalization of
propagation impairments,” IEEE Photonics Technology Letters., vol. 16, no. 2, pp. 674–676, 2004.
[4] R. I. K. P. B. Seb J. Savory, Giancarlo Gavioli, “Electronic compensation of chromatic dispersion using a digital
coherent receiver,” Optical express, vol. 15, no. 5, 2007.
[5] S. J. Savoryl, “Digital filters for coherent optical receivers,” Optical express, vol. 16, no. 2, 2008.
[6] G. Goldfarb and G. Li, “Chromatic dispersion compensation using digital iir filtering with coherent detection,” IEEE
Photonics Technology letters, vol. 19, no. 13, pp. 969–971, 2007.
[7] Jones calculus. http://en.wikipedia.org/wiki/Jones calculus.
[8] D. N. Gorald, “Self-recovering equalization and carrier tracking in two-dimensional data communication system,” IEEE
transactions on communication, vol. 28, no. 11, 1980.
[9] M. G. Taylor, “Phase estimation methods for optical coherent detection using digital signal processing,” Journal of
Lightwave Technology, vol. 27, no. 7, pp. 901–914, 2009.
[10] S. T. Kazuro Kikuchi, Dany-Sebastien Ly-Gagnon and K. Katoh, “Coherent detection of optical quadrature phase-shift
keying signals with carrier phase estimation,” Journal of Lightwave Technology, vol. 24, no. 1, pp. 12–21, 2006.
[11] K. K. Dany-Sebastien Ly-Gagnon and K. Kikuchi, “Unrepeated optical transmission of 2o=gbit/s quadrature phase-
shift keying signal over 210-km transmission using homodyne phase-diversity receiver and digital signal processing,”
Electronic letters, vol. 41, no. 4, 2005.
[12] C. R. S. F. et al., “Coherent equalization and plomux-rz-dqpsk for robust 100-ge transmission,” Journal of Lightwave
Technology, vol. 26, no. 1, pp. 64–71, 2008.