The objective of this research is to develop techniques for increasing transmission capacity in both short-reach and long-reach fiber networks. The approach is to use spatial multiplexing in multimode fiber, which is a form of multi-input, multi-output transmission.
Intellectual Merit: Short-reach fiber systems, such as data-center and local-area networks, already use multimode fiber with direct detection, but are limited by modal dispersion to only 10 Gbit/s per fiber. Adaptive optical signal processing will be used to prevent modal dispersion and enable spatial multiplexing in these systems. The approach developed in this project will simultaneously overcome modal dispersion and achieve multiplicative capacity gains in direct-detection systems.
Longer-reach systems currently use single-mode fiber, and are now using coherent detection, which preserves information on phase and polarization of optical signals. Adaptive digital signal processing techniques will be developed to compensate modal dispersion and enable spatial multiplexing. Mode-dependent fiber losses and amplifier gains can cause signal fading, as in wireless systems. System-level approaches, such as modal-frequency diversity and space-time coding, will ensure reliability without significantly sacrificing multiplexing gain. Fundamental diversity-multiplexing tradeoffs will be studied.
Broader Impacts: This project will enable dramatic increases in optical transmission capacity, and will help lower the cost per bit, benefiting research, education, healthcare, commerce, and security. New applications of multimode media will be enabled, complementing single-mode media. New connections between optical and wireless communications will be made, inspiring engineers in both fields. High-school students will be exposed to current research. Undergraduate and graduate students from diverse backgrounds will receive research training. Results of research will be integrated into undergraduate and graduate curricula.
Researchers at Stanford University have investigated mode-division multiplexing (MDM) in multi-mode fiber (MMF) for scaling optical networks to very high throughput, while reducing the transmission and switching cost per bit. Single-mode fiber (SMF) systems are now reaching capacity limits imposed by noise and nonlinearity. MDM exploits the spatial dimensions in MMF to increase the transmission capacity per fiber, yielding compelling advantages over transmission in parallel SMFs. By reducing the number of fibers required to accommodate a given traffic volume, MDM decreases the number of switches, amplifiers and other optical components required to handle a given traffic volume, reducing switching cost and simplifying network control. MDM enables enhanced component integration, increasing density and reducing transmission cost and energy consumption. The figure shows a long-haul MDM system using coherent detection and multi-input multi-output (MIMO) digital signal processing. As signals propagate in an MDM system, they are subject to crosstalk from mode coupling within the fiber, and modal dispersion (MD), a mode-dependent group velocity that is analogous to the multipath delay spread in wireless systems. In an MDM system, mode-dependent gain and loss (MDG) in amplifiers and fibers leads to signal power variations analogous to multipath fading in wireless systems. Mode coupling not only induces crosstalk between multiplexed signals, but also causes MD and MDG to become random, time-varying effects. The Stanford group derived the statistics describing MD and MDG in the strong-coupling regime relevant to long-haul MDM systems. The statistics of coupled MD govern the complexity of MIMO equalization. The statistics of coupled MDG represent fading distribution governing error probabilities, channel capacities and outage probabilities. The Stanford group studied the impact of MDG on MDM systems. MDG decreases average capacity and can cause outage. They showed that in a wideband system in which the signal bandwidth is much larger than the coherence bandwidth for MDG, frequency diversity can efficiently reduce outage probability, allowing outage capacity to approach average capacity, without requiring space-time codes or other methods that can reduce throughput or increase complexity. The Stanford group studied components and subsystems that are critical for enabling MDM systems. They optimized erbium-doped fiber amplifiers to minimize MDG, and devised adaptive optical equalizers for MDG, which are analogous to devices widely used to equalize wavelength-dependent gain in WDM systems. They studied fiber designs to minimize group delay spread, including a graded-index MMF with graded depressed cladding that achieved very low uncoupled delay spreads. They studied coupled-core multi-core fibers, showing they can achieve ultra-low delay spreads. MIMO equalization is potentially a major technical challenge for MDM systems, owing to the high MIMO dimensionality and potentially long delay spread. The Stanford group studied fixed and adaptive equalization methods. Their work showed that with careful optimization of equalizer architecture and control of system delay spread, the complexity per bit can be only modestly higher than in SMF, and increases slowly with the number of modes The project has enhanced understanding of how spatial dimensions can be exploited to increase optical network transmission capacity, while simplifying switching. Likewise, the project has enhanced understanding of MIMO optical channel statistics, how they depend on physical link design, and how they affect signal processing complexity and system performance. The channel statistics derived by the Stanford team are as fundamental to optical systems as the Rayleigh and Rician distributions are to wireless systems. The project has created and strengthened connections between the optical and wireless disciplines, inspiring engineers in both fields. By increasing throughput while reducing cost and energy per bit in optical networks, the project has helped enable the continued growth of information technologies that are indispensable to commerce, education, health care and homeland security. The project has provided advanced training of future engineers and educators–including those of diverse ethnicity and gender–in the interdisciplinary field of optical communications. Knowledge created in the project has been integrated into lecture and laboratory courses at Stanford. Outreach to high school and elementary school students has stimulated their interest in optics and careers in science and engineering. The Stanford team has met with economically disadvantaged and minority high school students visiting Stanford, and has students at an economically and ethnically diverse elementary school in San Francisco. In both venues, they offered hands-on optics experiments and encourage young students to aspire to technical careers.
