Defense Notices

The Doppler tolerance of a waveform refers to its behavior when subjected to a fast-time Doppler shift imposed by scattering that involves nonnegligible radial velocity. While previous efforts have established decision-based criteria that lead to a binary judgment of Doppler tolerant or intolerant, it is also useful to establish a measure of the degree of Doppler tolerance. The purpose in doing so is to introduce a Doppler "quasi-tolerant" trade-space that can ultimately inform automated/cognitive waveform design in increasingly complex and dynamic radio frequency (RF) environments. This idea of Doppler quasi-tolerance leads to the development of random FM (RFM) waveforms that retain a degree of Doppler tolerance while still providing the diversity of a nonrepeating waveform structure. The ensuing ambiguity functions split the delay/Doppler ridge into a variety of different patterns. Since these patterns are known at transmission, a strategy for appropriate coherent slow time combining is demonstrated in simulation. Separately, the application of slow-time coding (STC) to the Doppler-tolerant linear FM (LFM) waveform has been examined for disambiguation of multiple range ambiguities. However, using STC with non-adaptive Doppler processing often results in high Doppler "cross-ambiguity" side lobes that can hinder range disambiguation despite the degree of separability imparted by STC. To enhance this separability, a gradient-based optimization of STC sequences is developed, and a "multi-range" (MR) modification to the reiterative super-resolution (RISR) approach that accounts for the distinct range interval structures from STC is examined. The efficacy of these approaches is demonstrated using open-air measurements. Pulse agility is an alternative range disambiguation technique that relies on pulse-to-pulse waveform separability. Although pulse-agile waveforms are often uncorrelated and therefore amenable to range disambiguation, they may exhibit poor Doppler tolerance. To preserve Doppler tolerance and achieve separability, a class of hybrid waveforms is developed whereby a phase code is embedded on an LFM base waveform. A gradient-based optimization is developed for this waveform structure to achieve enhanced suppression of range-folded scattering in desired delay/Doppler regions. The Doppler tolerance and separability of the optimized waveforms are examined in simulation, and open-air measurements are used to demonstrate the range disambiguation capability.

The rapid growth of data traffic is pushing fiber-optic communication systems toward increasingly higher capacities. Meeting this demand requires advances across optical transmitters, in-line amplifiers, and usable gain bandwidth. This thesis investigates three key hardware technologies for next-generation, ultra-high-capacity optical networks: the multi-wavelength light source, the nonlinear amplifier, and the extended-band amplifier, each addressing a distinct bottleneck in scaling system capacity.                          

The first study investigates a single-section InGaAsP quantum-well Fabry--Perot laser diode as a compact, high-power, multi-carrier comb source with an output power exceeding 21 dBm. A single device generating multiple carriers can replace an array of discrete lasers, reducing cost and complexity in a coherent transmitter. The device is characterized in terms of its RF mixing linewidth, measured to be 20 kHz at full width at half maximum, relative intensity noise, and low phase noise beyond 1 GHz. Its suitability as a coherent transmitter is demonstrated using twenty 15 GBaud double sideband 16-QAM channels transmitted over 78.3 km of single-mode fiber. The system capability can reach beyond 4 Tb/s with a single laser in the transmitter using polarization multiplexing.                     

The second study demonstrates a polarization-insensitive fiber-optic parametric amplifier. The amplifier is based on a polarization-diverse bidirectional loop containing two equal-length, 100m highly nonlinear polarization-maintaining fibers connected through a 90° axis-rotation splice. This configuration eliminates the need for active polarization control, compensates for polarization-mode-dispersion effects, and suppresses stimulated Brillouin scattering between the counter-propagating pumps at pump powers of up to 30 dBm. When operated as a preamplifier in a 400 Gb/s dual-polarization 16-QAM transmission system, the amplifier achieves a noise figure below 3 dB, confirming its viability for practical, polarization-diverse signals.                      

The third study presents an experimental characterization of a 468m Bismuth-doped Germanosilicate fiber operating in the E and S band. The amplifier can provide a gain up to 33 dB when operated at a pump power of 23 dBm. A bidirectional rate-equation model of the measured gain and ASE spectra resolves the contributions of the individual active-center subtypes to net gain. The model further quantifies the impact of Rayleigh backscattering in long bismuth-doped fibers. Together, these results indicate that long fibers operate more efficiently in the longer-wavelength region. Combined with a conventional Erbium-doped amplifier, such a Bismuth-doped stage could extend gain coverage across the S through L bands, making it a single extended-band amplifier.                        

Together, these studies address three complementary elements of high-capacity optical communication systems: multi-channel source generation, nonlinear parametric amplification, and extended-band doped-fiber amplification. By examining the physical interactions, noise analysis, and system-level impairments associated with each technology, this thesis addresses several of the hardware technologies needed to overcome major capacity-scaling limitations of current optical networks.