Optical Communications

While optical communications was closely related to long-haul transmission systems in the past, more than 50 % of the global data traffic is attributed to datacenter interconnects nowadays. For energy-efficient and fast communication between the servers and clients, fiber-optic solutions need to replace electrical connections and shrink down in size. As a result, the fiber-optics market at present is mainly driven by the development of photonic-integrated circuits and high-capacity short-reach transmission systems.

Under the lead of Prof. Randel, the Optical Communications Group is working towards data transmission beyond data rates of 1 Tbit/s exploiting photonic-assisted generation of high symbol rates as well as coherent modulation formats with high spectral efficiency. Since high-throughput transceivers are susceptible to noise, transmitter impairments and nonlinear effects, their optimization requires merging knowledge from a variety of research areas like coding, signal processing, and nonlinear optics. Following a recent trend in the communications community, researchers at IPQ explore whether machine learning can be employed for optimization of such systems. Novel DSP algorithms developed at IPQ are not only tested on offline waveforms, but also in real-time systems including field programmable gate array (FPGA) platforms. This is crucial for a reliable estimate of applicability, latency, and computational complexity of these digital techniques.

Another research topic is the investigation of coherent architectures for passive optical networks (PONs) that provide broadband access to end customers and usually span a few kilometers. Such networks feature a point-to-multipoint topology and require low-cost hardware at the customer side that is frequency synchronized with the optical line terminal. Researchers at IPQ examine how frequency combs can help achieving higher data rates with cheap lasers that tend to drift over a large wavelength range. The project KIGLIS targets on the optimization of PONs with respect to performance, reliability, cost efficiency, and energy efficiency by means of artificial intelligence (AI). To this end, the project focusses on the application scenario of a “smart city” and supports the idea of fixed mobile convergence, i.e., a seamless connection between fiber-optical communication systems and local access wireless technologies such as 5G. In a field trial, the potential benefits of AI in handling the huge amount of sensor data in the context of autonomous driving shall be investigated.

Frequency combs are also an essential part of the STARFALL project that deals with the development of novel transceiver architectures that support fiber-optic spatial division multiplexing (SDM). SDM is an attractive complement to the existing wavelength division multiplexing (WDM) systems because it allows scaling the channel capacity linearly with the number of spatial paths. Challenges impeding commercialization are power consumption, hardware complexity, and DSP complexity, which all are addressed within STARFALL. For shrinking the size of SDM transceivers, the project KONFORM studies ultra-compact 3D multi-mode converters that can be printed by a two-photon lithography process.

In close collaboration with the Teratronics Signal Processing Group of Prof. Koos, the Optical Communications Group also advances high-capacity wireless transmission systems in the terahertz (THz) or optical regime. Recently, the group achieved some groundbreaking system demonstrations that may pave the way towards high-speed point-to-point links between base stations in urban or peripheral areas, where fiber deployment would cause massive costs. This work is continued as part of the Open6G Hub Germany. Free-space optical transmission systems are an alternative to THz systems when larger distances need to be bridged. However, optical signals are very susceptible to adverse weather conditions and turbulent atmosphere. At IPQ, researchers explore to which extend the effect of turbulence can be digitally mitigated at the receiver side of the transmission link.

Over the past decade, development in coherent transceiver technologies has redefined our ability to analyze, synthesize, and manipulate optical fields across time, frequency, quadrature, and polarization. Signal processing functions that seemed to be realizable only in the optical domain due to electronic bandwidth limitations and DSP constraints, can now also be realized by high-performance digital processing of the optical field after coherent reception. An interesting example of that paradigm shift is an optic-electronic-optic (OEO) interferometer, which can potentially extend the functionality of conventional interferometers towards arbitrary linear, non-linear, and even time-variable functions. Such a device would open up new possibilities for network routing. The aim of the project INTERFERE is a first experimental demonstration of an OEO interferometer based on an integrated optics framework.

For system-level experiments, the group can resort to one of the best-equipped optical transmission laboratories worldwide. It comprises multi-format transmitters to generate high-speed electrical data that is modulated onto optical carriers using either commercial Lithium Niobate or in-house designed modulators, (for details read more about the Hybrid Photonic Integrated Circuit Group). For reception, dual-polarization integrated coherent receivers or high-speed photodiodes can be combined with real-time oscilloscopes capturing frequencies up to 100 GHz. For transmission experiments, the lab is equipped with reconfigurable optical switches, multiplexers, and hundreds of kilometers of single-mode fiber (SMF).