The strongest gravitational waves, those from astrophysical sources such as
neutron star binary coalescence, supernovae, or black hole formation, are
expected to produce strains of 
or smaller. This means that
gravitational wave detectors must meet daunting performance specifications
if success is to be achieved. Simply to register such small spatial
perturbations requires measurements of unprecedented precision.
Simultaneously, the test masses, which must be free to respond to the
gravitational wave, must also be isolated to an unprecendented degree from
other disturbing influences.
Measurement precision will be achieved by making the arms of the Michelson interferometer from Fabry-Perot cavities (allowing the signal to accumulate for a time comparable to half the period of the wave), and by illuminating the interferometer with sufficient optical power to make the fundamental limit to strain measurement precision, shot noise in the power measurement at the interferometer output port, small enough.
The two chief disturbances that compete with the gravitational wave effect are seismic noise and thermal noise. External vibrations from natural or man-made sources will be isolated by mechanical filters of many poles, including the pendulum suspensions of the test masses themselves. Minimization of thermal noise requires designs that reduce to the lowest possible levels the amount of mechanical dissipation in the test masses themselves, and in their suspensions.
Estimates of the noise spectra of interferometric gravitational wave detectors follow a basic pattern: the high frequency band will be dominated by the shot noise in the precision of the interferometric readout, while the low frequency band will be dominated by insufficiently filtered seismic noise. In between, thermal noise is likely to be the strongest source of noise. In the earliest versions of these devices, seismic noise and shot noise may be high enough that thermal noise will dominate over only a very small band. An estimate of the noise budget for the initial LIGO interferometer is shown in Figure 1.
There are well-developed strategies for reducing shot noise (with high power lasers and various forms of the optical techniques known as recycling), and for reducing seismic noise (by constructing vibration filters with lower resonant frequencies.) Unless there is comparable progress in reducing thermal noise from the levels we know how to achieve now, future improvements in gravitational wave sensitivity may come slowly. Without such progress, sensitivity in the entire band from 10 Hz to 1 kHz may be limited by thermal noise. The goal of our research program at Syracuse University is to contribute to a balanced international effort to improve the sensitivity of gravitational wave interferometers, by searching for ways to minimize thermal noise.