Research Interests

A fundamental unsolved problem in fiber optic communication systems is the question of ultimate channel capacity. While the transmission capacity of commercial fiber optic communication systems based on Wavelength Division Multiplexing (WDM) continues to double roughly every 1-2 years (today’s standard is ~ 100 channels @ 10Gbps per channel), it is not known how far this scaling in capacity can continue. A related unsolved problem is the maximum transmission distance, given a certain channel capacity, without resorting to optical-electrical-optical (OEO) regeneration. These problems have not been solved for light wave communications, in contrast to say RF communications, because optical fiber is a nonlinear medium. So for example, many important results in communications theory, such as Shannon’s Capacity Theorem, cannot be applied to Ultra-Dense WDM (UDWDM), where fiber nonlinear effects dominate. The key to solving the problem of ultimate channel capacity is to improve our understanding of how nonlinear interactions degrade the performance of fiber optic communication systems, and to apply this knowledge to design optimum transmitters, optical amplifiers, filters, and receivers to attain the theoretical maximum (currently unknown) channel capacity.

So far, engineers have been able to increase channel capacity by utilizing more of the available low loss window in fiber (e.g. developing C+L-Band Erbium doped fiber amplifier, and Raman amplifier), and increasing spectral efficiency (i.e. channel bit rate/ channel spacing), without running into any fundamental physical barriers. The current industry benchmark for channel spacing in UDWDM is 25 GHz for 10Gbps channels, and 100GHz for 40Gbps channels, which gives a spectral efficiency of 40%. Research is ongoing to push spectral efficiency toward 80%, with 12.5 GHz channel spacing at 10 Gbps (50GHz channel spacing at 40 Gbsp). While there are many practical problems that need to be solved in UDWDM, such as improving the wavelength stability of DFB lasers (< 100 MHz variation), and developing ultra-narrow (high-Q) optical filters with low dispersion (< 10 ps delay variation), the hardest problem to overcome is the substantial nonlinear interaction between channels at UDWDM channel spacing, which is expected to form a fundamental physical barrier as the spectral efficiency approaches 100% and beyond.

There are several interesting problems in UDWDM that need to be addressed. These include:

1. Rigorous theory of channel capacity.

2. Scaling of fiber nonlinear impairments with channel spacing, bit rate, and transmission distance.

3. Optimum modulation formats.

4. Optical filter and receiver design.

5. Laser wavelength-locking techniques.

6. Optical/electrical equalization.