Over the past ten years the fiber optic gyro has moved from an initial laboratory demonstration in 1976 to commercially viable production prototypes in 1986. While detailed reviews of progress made by a number of organizations are contained in the proceedings this paper provides a brief overview of the field.
This paper is a brief review of the history of fiberoptic gyros with emphasis on some of the key concepts that were critical to the development of this field. The contributions that M.I.T. may have made to this field are mentioned in this context. This review is by no means exhaustive and apologies are offered to those contributors that may have been overlooked.
McDonnell Douglas initiated a program to develop fiber optic gyros in 1977. During the 1977 to 1986 time period a series of key milestones were achieved including the first fiber optic gyro with linear digital output, the first all solid-state optical gyro, the first fiber optic gyro optical head of a size compatible with tactical requirements, the first practical means of correcting for scale factor error due to wavelength shifts and the first fiber optic gyro product. These efforts as well as the future direction of fiber optic gyro development at McDonnell Douglas are summarized.
Fiber optic gyro development began, approximately ten years ago with Vali and Shorthill1. Progress has been rapid over these years due to the efforts of a large number of scientists and engineers around the world. Optical noise sources have been identified and reduced. Noise performance of present day fiber gyros is essentially set by shot noise or electronics noise. Many sources of optical bias errors have also been identified and reduced. Currently, scale factor errors, packaging, and environmental ruggedness are being addressed along with cost, reliability, and production issues to turn the fiber gyro from a labora-tory instrument into a product.
Based on phase and frequency modulation techniques two types of fiber-optic gyroscopes are under development at SEL for different applications. A set of identical prototype units has been built for environmental testing and for system concept evaluation. The experience made with the manufacturing of the devices is discussed, regarding failures, performance limits and further improvements. The critical components in the optical part of the gyros are defined.
In the first period of the fiber-gyro development fundamental problems such as signal noise and bias stability (drift) had to be solved. By the use of light with short coherence length disturbing interferences were avoided and a noise equivalent rotation rate of about 0,1°/h (Bandwith 1 Hz) corresponding to a random walk of 1.2 10-3°/h was achieved. The reciprocal arrangement of the Sagnac-interferometer the so called 'minimum configuration' (Fig. 1) yield a bias stability of better than 0.1°/h. These data are sufficient for many applications of the fiber-gyro. Another important parameter is the scale factor error. There are only few reports on scale factor measurements. Up to now an accuracy in the range of 10-2-10-3 was achieved.
A theoretical analysis of the non-reciprocal phase error in a highly-birefringent fiber gyroscope is presented, characterizing the different types of errors and their sources in a general form. These error terms are due to polarization cross coupling in the fiber and in the fiber components and can be reduced by modulating the magnitude of the fiber birefringence at appropriate locations in the optical circuit and by balancing the optical power between the two polarization eigenmodes of the fiber at the input of the interferometer. Experimental results demonstrating these techniques show a 20 dB relief on the performance requirements of the polarizer usually employed for drift reduction.
LSI has been involved in the development of optical gyros for 17 years. Early work 1 (1969 to 1975) included a passive gyro approach that measured a change in polarization state as a function of Sagnac phase. From 1979 to the present, LSI has been actively developing a passive optical rate sensor using fiber and integrated optics. The present sensor design is a closed loop interferometer approach designed to be insensitive to changes in source intensity and wavelength.
This paper describes the optomechanical systems incorporated into the optical module of the breadboard closed loop fiber optic gyroscope. The performance achieved was a rotation rate noise of 5°/hr for 1 second in-tegration, 0.55°/hr for 100 second integration and extrapolated drift of 0.09Whr. The long term (6 hours) bias stability was 1°/hr. A detailed description is directed towards optomechanical systems and related mechanical problems. The results of the investigation into the fiber sensing coil designs and the response of various coil spool geometry configurations to thermal gradients and transients are reported. Also, a sample gyroscope drift run is presented.
This paper presents the results of a series of experiments performed on a fibre optic gyroscope to determine the effect of different phase modulation frequencies and modulation depths on coherent backscatter noise. The results suggest that coherent backscatter is modulated in quadrature with the Sagnac signal although a presently unexplained noise source is also detected in-phase with the Sagnac signal.
The development of all-fiber-optic gyroscopes at Stanford University is reviewed. Fiber-optic components for the gyroscopes are described. Techniques to suppress various error sources, and different approaches to a linear scale factor with wide dynamic range, are discussed. Fiber-optic single sideband frequency shifters for closed-loop operation of gyroscopes are described. A new approach to high birefringence fiber gyroscopes, with broadband optical source and birefringence modulation, is presented. Other approaches, including the pulsed reentrant fiber gyro, are shown.
The test results of a high performance, open-loop, all-fiber-optic rotation sensor are reported. The device used a low-coherence superradiant diode as the optical source with a polarization-maintaining sensing coil, polarizer, and loop coupler. Various contributors to bias drift that were systematically analyzed and measured are discussed. Various noise sources contributing to the random walk coefficient are described. The thermal sensitivity of the sensor is described.
The design of in-line phase and frequency shifters for single mode fibers is discussed. The basic configuration consists of a cylindrically symmetric piezoelectric transducer fabricated on the fiber surface. For the phase shifter the fiber core is concentric and no polarization coupling occurs. The frequency shifter utilizes a birefringent fiber with a non-concentric core and transducer pairs driven 90° out of phase. Factors affecting the performance of both devices are discussed.
An all fiber single-sideband frequency shifter using acousto-optic interaction in a double-moded fiber has been developed. Conversion efficiency of 100% with 35 dB suppression of the carrier and image sideband has been achieved at a frequency of 8 MHz using an input electrical power of 100mW.
Integrated optics is appearing as a practical solution for the fiber gyroscope. We describe quarter of a liter and tenth of a liter brassboards using such a multifunction circuit (in a so-called "Y-tap" configuration). Random walk performance of 2 x 10-°/√h has been demonstrated as long term bias stability of ± 0.3°/h over a 15°C temperature range.
A low-level, but continuing, fiber-gyro development activity has been carried on at the Jet Propulsion Laboratory since 1977. The activity was originated because of a recognition of the potential for low-cost, high-performance gyros suitable for interplanetary spacecraft. An early decision was made to concentrate available resources on supporting the development of electro-optically active channel waveguide components which could be fabricated by mask diffusion processes. Titanium-indiffused lithium niobate waveguide components used at 0.83 μm wavelength were first tested and then abandoned because of instabilities caused by so-called optical damage. Components fabricated for use at 1.3 μm wavelength have proven to be stable. A gyro configuration concept based upon 1.3 μm channel waveguide components has evolved, and a baseline 1.3 μm all-fiber gyro has been assembled and tested.
The Jet Propulsion Laboratory is developing a Fiber Optic Rotation Sensor (FORS) for use on the Mariner Mark II series of planetary explorer craft and in other space applications. FORS is a closed-loop, phase-nulling device and embodies a number of interesting innovations. Chief among these are the incorporation of the device's couplers, phase modulators and polarizer on a single lithium niobate (LiNbO3) integrate optics chip and a novel means of reading out angular position and rotation rate based on optical beat detection Various aspects of the FORS design and operation are described and discussed. Particular attention is paid to analyzing errors attributable to polarizer imperfection and the so-called residual Michelson effect.
An optical fiber gyroscope configuration which utilizes dual wavelength determination of the Sagnac shift is described. It is shown that this provides a means of extending the unambiguous rotation sensing range, and thus the dynamic range, of the gyroscope. An increase in unambiguous sensing range ~ x 45 is demonstrated using an all-fiber gyroscope, giving an absolute dynamic range of ~ 10 6.
Progress in developing serrodyne frequency shifters for fiber gyroscopes is reviewed. The objective of this work has been to maximize sideband suppression and minimize the effects of small variations in input polarization on device performance. The LiNb03 device described here utilizes a specialized sawtooth drive circuit and single-polarization wave-guides to achieve 41-dB sideband suppression.
Integrated optics will play a major role in the next generation of fiber gyros since the potential exists for mass producing low-cost integrated gyro front-end circuits. All the necessary components (phase modulators, polarizers, and power splitters) have been demonstrated with Ti:LiNb03 waveguide technology. An attractive alternative of combining the 3 dB splitter and polarizer functions in a 3x3 directional coupler is presented.
The Phase-Nulling Optical Gyro (PNOG) has been under development since 1977. Techniques that have been developed to refine the PNOG can be adapted to develop fiber-optic sensors to measure other effects such as acoustic waves, electric and magnetic fields, pressure, temperature and strain. As an example of utilizing PNOG technology to develop other types of fiber-optic sensors, a Sagnac based strain sensor will be described. Emphasis will be placed on techniques and hardware which have been derived from PNOG development. A laboratory model will then be described and data taken with this device will be presented.
The Sagnac phase shift between two counterpropagating beams is the basis for all optical gyroscopes [11, although it is detected in a variety of ways. In the ring laser gyro and various "closed loop" optical gyroscopes, the scale factor (e.g. counts per unit rotation rate) is fixed by the area of the optical medium whereas the phase reading fiber optic gyroscope (FOG) has a scale factor which increases with increasing length of optical fiber. Thus development of low-loss fibers holds the possibility of extremely sensitive gyroscope operation, and has made the FOG a competitive optical gyroscope. Although most of the first decade of FOG development has concentrated on improving sensitivity, attention is now turning to other aspects of performance.
Since 83 LITEF is investigating different concepts of fiber optic gyroscopes theoretically anu experimentally. It started with the passive resonator gyro which promises a good scale factor linearity and repeatability. For building labmodels it was necessary to develop single-mode fiber optic components, especially low-loss couplers. For comparison Sagnac interferometer gyros were built, which generally showed better bias stability than the resonator gyros.
Theoretical and experimental work at The Charles Stark Draper Laboratory, Inc. (CSDL) indicates that it is feasible and very attractive to operate the fiber optic gyro in the passive resonant cavity mode. Recent advances in the coherent communications field indicate promise for the ready availability in the near future of narrow line diode lasers, single polarization fibers, and electro-optical waveguide signal processing circuits at the 1.3 and 1.55 micron wavelengths. The purpose of this paper is to describe the operation, characteristics, and performance of a closed loop fiber optic resonator gyro and compare it to the interferometric fiber optic gyro. The limitations and potential of the passive resonator gyro will be discussed.
A novel optical fibre resonator rotation sensor using a low coherence source is described. The sensitivity of the system is compared to the conventional phase modulated Sagnac interferometer and a higher sensitivity appears feasible.
Fiber optic gyroscope (FOG) development current status is reviewed and summarized. These FOG categories are reviewed:
•fiber - plus - bulk optics
•Fiber - plus IOC US development funding of these gyros, segmented into internal R&D and government funding, is forecasted, 1986-1996. The US production of fiber optic gyroscopes is forecasted, 1986-1996. The forecast is segmented into repeatibility rate categories and related applications, and includes both quantity and value data. Component price trends are forecasted. The trend to operation at 1.3-1.6 microns is forecasted.
High performance fiber optic gyroscopes require a high degree of scale factor stability. Scale factor is directly dependent' on the wavelength of the light source, regardless of whether the fiber optic gyroscope is operated open or closed loop. The light source of choice is a broadband AlGaAs superluminescent diode (SLD). This paper describes a compact, stable, passive SLD wavelength monitoring device which can measure the mean wavelength of an SLD to an accuracy that meets stability requirements.
Reduced cost, greater reliability, and compatibility with integrated optics are some of the advantages gained by incorporating semiconductor lasers into the design of passive resonant ring laser gyroscopes (PRRLG). Unfortunately, the large spectral linewidth typical of semiconductor lasers greatly limits the sensitivity of such a configuration. This research details the modification of a laser diode to obtain a stable, single mode, narrow spectral output using an optical feedback configuration compatible with the input optics to a PRRLG. Spectral narrowing and modal dynamics of the external coupled cavity laser were measured. By temperature stabilization and optical feedback a free running linewidth of 60 MHz was reduced to approximately 13 MHz or less and was maintained in a single mode for over 20 minutes.
High power superluminescent diodes(SLD)havebeen fabricated from both gain guided and index guided AlGaAs/GaAs lasers. The spectral width (FWHM) of 15nm for the gain guided SLD and 8 nm for the index guided SLD are typical measured values. Output power of over 30mW/ facet is achieved for the SLD at wavelength of .83um. Over 3 mW has been obtained from the core of pigtailed single mode polarization preserving fiber. SLDsoperating at 1.3 um wave-length region has also been fabricated from index guided GaInAsP/InP lasers. Output power of 15 mW have been obtained. Coupling into single mode polarization preserving fiber with output power from fiber core over 2 mW has been achieved. Life testing of both SLDson heat sinks and single mode fiber pigtailed SLD5havebeen performed at different temperatures. Room temperature extrapolated life time of over 100,000 hours is expected.
Three modifications of the twin channel laser structure have been fabricated and characterized for continuous wave power and spectral limitations in superluminiscent diode (SLD) behavior. The conventional antireflective coating technique produced lasing - limited power less than 10mW. Fabry-Perot spectral modulation was typically greater than 50 percent at 7mW. The angle stripe approach yielded thermal - rollover limited powers of 30mW. Spectral halfwidth of 160 Å and 5% Fabry-Perot modulation were typical. The truncated stripe structure produced 15mW.
Extinction measurements for both polarization maintaining and polarizing fibers as a function of length are reported. For fiber lengths of more than 10 meters, the extinction coefficient of polarizing fiber exceeds that of polarization maintaining fiber.
Polarization maintaining fiber with a near rectangular cross section, made by preform deformation, has demonstrated extinction ratios for 5 meter lengths of 43 dB, h parameter values of 1.7x10-6 m-1 for 100 meter lengths and losses of 0.8 dB/km at 1.3 and 1.5ttm. Polarizing fiber with a W type index profile, also made by preform deformation, has demonstrated extinction ratios of greater than 30 dB in straight lengths as short as 5 cm. with orthogonal polarization cutoff wavelengths separated by 0.1μm and insertion loss of less than 3 dB. The rectangular shape of these fibers facilitates location of the principal axes and allows winding such that a principal axes can be aligned with the bending plane.
Polarization retaining single-mode fibers are being developed using the stress-rod approach. Design parameter values have been varied to optimize birefringence, polarization holding, and low attenuation rates. Test configurations have been developed to characterize the fibers under simulated deployment bending conditions. The results show that good polarization and attenuation properties can be maintained for long lengths of fibers wound into compact sensor configurations.
This bibliography, numbering about 410 items, is intended to be as complete as practicable for work published in journals and conference proceedings, technical reports, dissertations, or other forms that are readily available to the public. All entries relate directly to actual fiber-optic gyros and rotation sensing, and include theory, analysis, experiments, and engineering development. Work in supporting technologies, such as fiber, light sources, and components is not included unless directed specifically to gyro issues. A few non-gyro items are included, which address fundamental effects that are important to gyro operation or that may limit performance. Patents are not included. The bibliography was accumulated with little effort over a period of several years using a personal computer. Latest additions were made late August 1986. The bibliography should serve as a useful reference tool for all workers. And, excepting a significant amount of yet unpublished corporate work, it provides an excellent historical overview in keeping with the theme of this conference.