Improving fiber gyroscope performance using air-core fiber
Fiber-optic gyroscopes (FOGs) have proven to be one of the most successful fiber sensor technologies, with effective deployment in a variety of commercial and military applications. Nonetheless, current FOGs are limited by several sources of error. These include excess source noise from broadband light and fluctuations in the source mean wavelength, resulting in instability in the FOG scale factor (which relates the sensor signal to the rotation rate of the fiber loop). An additional problem is long-term drift of the FOG output due to thermal variations in the coil, known as the Shupe effect. It is interesting to note that broadband, incoherent light sources contribute to two of these errors, and reducing them to meet the stringent requirements of applications such as inertial navigation for airplane guidance has proven challenging. Here, we describe our recent work on the development of a new fiber gyroscope configuration designed to overcome these problems.
Figure 1 shows our proposed modifications in schematic fashion: replacing the solid-core fiber loop with a hollow-core photonic-bandgap fiber (PBF), and using a narrow-linewidth laser instead of the traditional broadband light source. We predicted that such basic changes to the fiber gyro would have several key benefits. In a hollow-core PBF, the fundamental mode travels mostly in air. Because the refractive index of air has a much weaker temperature dependence than that of silica, long-term drift due to the Shupe effect should be reduced. Since air has a much weaker Kerr nonlinearity than silica, any Kerr-induced nonreciprocity in the FOG will also be greatly mitigated. Similarly, since air scatters much less than silica, coherent backscatter in a PBF FOG should ultimately be weaker than in a traditional FOG. Early FOG research abandoned narrow-linewidth sources in favor of broadband ones precisely to reduce the large Kerr-induced drift and high coherent backscattering noise present when the FOG is probed with a laser. When the sensing coil is made of a hollow-core fiber, it should be possible to interrogate the FOG using a highly coherent laser source, and consequently eliminate excess noise and improve mean wavelength stability.
Several of these improvements have been demonstrated experimentally. In a PBF FOG prototype operated with a narrowband laser, we measured a 6.5-fold reduction in thermal drift.1 The same gyro also showed a 250-fold lessening of the effective Kerr constant as compared to a conventional FOG.2 This has led to the demonstration of a laser-driven air-core FOG with an inferred Kerr-induced long-term drift low enough to meet the requirement for a 10-hour flight.
As reported elsewhere,3 we recently developed a new theory for predicting the coherent backscattering noise in fiber gyros. Although Rayleigh scattering in hollow-core PBFs is negligible, scattering in current PBFs is dominated by random perturbations of the core geometry. This induces higher backscattering than in typical silica fibers, although it is expected to decline to a level lower than that of a conventional fiber as manufacturing techniques improve. Even in the presence of backscattering, our modeling predicts that using very narrow laser linewidths (~1–10kHz) should reduce the total backscattering noise below the excess noise of a broadband source for a low-loss PBF coil. This is because backscattering noise in FOGs arises from the interaction of stationary, time-independent scatterers in the loop with the time-varying, random phase fluctuations of the source. Reducing fluctuations in the source phase leads to fewer fluctuations in both the backscattered signal and the gyro output. Figure 2 shows the expected backscattering noise predicted by our theory as well as experimental measurements performed using a 235m solid-core fiber coil. Decreasing the laser linewidth from 200 to 15kHz reduces the noise in the gyro by a factor of 2.4 (theoretically) and 1.4 (in practice). Further diminution is expected with narrower linewidths.
We are currently working on producing similar results using a PBF sensing coil.4 We anticipate that a similar reduction in backscattering noise will be possible, with the added benefits of PBFs mentioned earlier. This approach is ultimately expected to produce fiber-optic gyroscopes with short- and long-term noise performance sufficiently improved for more demanding applications in inertial navigation.
