Molly Piels

Machine Learning Techniques in Optical Communication

Machine learning techniques relevant for nonlinearity mitigation, carrier recovery, and nanoscale device characterization are reviewed and employed. Markov Chain Monte Carlo in combination with Bayesian filtering is employed within the nonlinear state-space framework and demonstrated for parameter estimation. It is shown that the time-varying effects of cross-phase modulation (XPM) induced polarization scattering and phase noise can be formulated within the nonlinear state-space model (SSM). This allows for tracking and compensation of the XPM induced impairments by employing approximate stochastic filtering methods such as extended Kalman or particle filtering. The achievable gains are dependent on the autocorrelation (AC) function properties of the impairments under consideration which is strongly dependent on the transmissions scenario. The gain of the compensation method are therefore investigated by varying the parameters of the AC function describing XPM-induced polarization scattering and phase noise. It is shown that an increase in the nonlinear tolerance of more than 2 dB is achievable for 32 Gbaud QPSK and 16-quadratic-amplitude modulation (QAM). It is also reviewed how laser rate equations can be formulated within the nonlinear state-space framework which allows for tracking of nonLorentzian laser phase noise lineshapes. It is experimentally demonstrated for 28 Gbaud 16-QAM signals that if the laser phase noise shape strongly deviates from the Lorentzian, phase noise tracking algorithms employing rate equation-based SSM result in a significant performance improvement (

Application of Machine Learning Techniques for Amplitude and Phase Noise Characterization

>

Machine Learning Techniques for Optical Performance Monitoring From Directly Detected PDM-QAM Signals

8 dB) compared to traditional approaches using digital phase-locked loop. Finally, Gaussian mixture model is reviewed and employed for nonlinear phase noise compensation and characterization of nanoscale devices structure variations.

Quaternary Polarization-Multiplexed Subsystem for High-Capacity IM/DD Optical Data Links

In this paper, tools from machine learning community, such as Bayesian filtering and expectation maximization parameter estimation, are presented and employed for laser amplitude and phase noise characterization. We show that phase noise estimation based on Bayesian filtering outperforms conventional time-domain approach in the presence of moderate measurement noise. Additionally, carrier synchronization based on Bayesian filtering, in combination with expectation maximization, is demonstrated for the first time experimentally.

400-GHz Wireless Transmission of 60-Gb/s Nyquist-QPSK Signals Using UTC-PD and Heterodyne Mixer

Linear signal processing algorithms are effective in dealing with linear transmission channel and linear signal detection, whereas the nonlinear signal processing algorithms, from the machine learning community, are effective in dealing with nonlinear transmission channel and nonlinear signal detection. In this paper, a brief overview of the various machine learning methods and their application in optical communication is presented and discussed. Moreover, supervised machine learning methods, such as neural networks and support vector machine, are experimentally demonstrated for in-band optical signal to noise ratio estimation and modulation format classification, respectively. The proposed methods accurately evaluate optical signals employing up to 64 quadrature amplitude modulation, at 32 Gbd, using only directly detected data.

60 Gbit/s 400 GHz wireless transmission

We demonstrate for the first time an intensity-modulated direct-detection link using four states of polarization. The four data-independent tributaries are each assigned distinct states of polarization to enable the receiver to separate the signals. Polarization rotation due to propagation over optical fiber is tracked and compensated with simple digital signal processing in Stokes space. Transmission below the forward error correction limit is shown for maximum net bitrates of 100 Gb/s (4 × 27 GBd) and 120 Gb/s (4 × 32 GBd) over 2-km standard single-mode fiber at a center wavelength of 1550 nm.

Bayesian filtering for phase noise characterization and carrier synchronization of up to 192 Gb/s PDM 64-QAM

We experimentally demonstrate an optical network compatible high-speed optoelectronics THz wireless transmission system operating at 400-GHz band. In the experiment, optical Nyquist quadrature phase-shift keying signals in a 12.5-GHz ultradense wavelength-division multiplexing grid is converted to the THz wireless radiation by photomixing in an antenna integrated unitravelling photodiode. The photomixing is transparent to optical modulation formats. We also demonstrate in the experiment the scalability of our system by applying single to four channels, as well as mixed three channels. Wireless transmission of a capacity of 60 Gb/s for four channels (15 Gb/s per channel) at 400-GHz band is successfully achieved, which pushes the data rates enabled by optoelectronics approach beyond the envelope in the frequency range above 300 GHz. Besides those, this study also validates the potential of bridging next generation 100 Gigabit Ethernet wired data stream for very high data rate indoor applications.

Ultrahigh-speed Si-integrated on-chip laser with tailored dynamic characteristics

We experimentally demonstrate a 400 GHz carrier wireless transmission system with real-time capable detection and demonstrate transmission of a 60 Gbit/s signal derived from optical Nyquist channels in a 12.5 GHz ultra-dense wavelength division multiplexing (UD-WDM) grid and carrying QPSK modulation. This is the highest data rate demonstrated for carrier frequencies above 300 GHz and also validates the feasibility of bridging between next generation 100 GbE wired data streams and indoor wireless applications.

Quad-polarization transmission for high-capacity IM/DD links

We show that phase noise estimation based on Bayesian filtering outperforms conventional time-domain approaches in the presence of moderate measurement noise. Additionally, carrier synchronization based on Bayesian filtering, in combination with expectation maximization, is demonstrated for the first time experimentally.

Compact silicon multimode waveguide spectrometer with enhanced bandwidth

For on-chip interconnects, an ideal light source should have an ultralow energy consumption per bandwidth (operating en-ergy) as well as sufficient output power for error-free detection. Nanocavity lasers have been considered the most ideal for smaller operating energy. However, they have a challenge in obtaining a sufficient output power. Here, as an alternative, we propose an ultrahigh-speed microcavity laser structure, based on a vertical cavity with a high-contrast grating (HCG) mirror for transverse magnetic (TM) polarisation. By using the TM HCG, a very small mode volume and an un-pumped compact optical feedback structure can be realised, which together tailor the frequency response function for achieving a very high speed at low injection currents. Furthermore, light can be emitted laterally into a Si waveguide. From an 1.54-μm optically-pumped laser, a 3-dB frequency of 27 GHz was obtained at a pumping level corresponding to sub-mA. Using measured 3-dB frequen-cies and calculated equivalent currents, the modulation current efficiency factor (MCEF) is estimated to be 42.1 GHz/mA1/2, which is superior among microcavity lasers. This shows a high potential for a very high speed at low injection currents or avery small heat generation at high bitrates, which are highly desirable for both on-chip and off-chip applications.