It is pointed out that the performance of speckle imaging or optical interferometer systems increases with (r sub 0/D) exp n, where r sub 0 is the atmospheric coherence length, D is the aperture size, and n is between 2 and 4. It has been determined that, since r sub 0 is about 10 cm at visible wavelengths and D may be several meters, selecting a site with a large r sub 0 becomes critical for 30-100-m baseline systems. A unique problem for such optical systems is the need for a relatively large, flat, approximately 100-m site; however, this is inconsistent with the atmospheric dynamics that produce optical sites. Albuquerque and Chilao Flats results indicate that katabatic flows produce r sub 0 values of 30-50 mm; on the other hand, large mountain tops tend to have large 50-200 m inner layers, making r sub 0 extremely sensitive to the surface heat flux and wind speed. It is concluded that few locations can achieve this; those along the California Pacific Coast and Mauna Kea are two such regions.

The U.S. Naval Observatory (USNO) is in the design and very early construction phases of a project to construct and operate an optical interferometer dedicated to astrometry. An overall description of the instrument is given with emphasis on those features which make the device uniquely suited for astrometric observations of the type envisioned. A brief discussion of these types of observations is included. A few of the more general considerations in the design and operation of the instrument are also mentioned. This instrument is in many ways an extension and an enlargement of the MARK III Interferometer at Mt. Wilson which has successfully made astrometric measurements.

A feasibility study of a multiple telescope array for high-spatial-resolution astronomy has been completed, and an initial design concept has been defined. The array (referred to as the CHARA array for Georgia State University's Center for High Angular Resolution Astronomy) would consist of seven 1-m-aperture telescopes in a VLA-type configuration contained within a circle of 400-m-diameter to provide a limiting resolution of 0.3 milliarcsec for stellar angular diameter measurements or 0.1 mas for binary-star measurements. The initial scientific program will be directed at the imaging of stars to determine stellar radii, masses, temperatures, distances, and surface morphology. The array is also intended to provide the means of developing techniques for the very-high-resolution imaging of a large class of objects with geometries far more complicated than those of stars and star systems.

Construction of a two-telescope Michelson spatial inteferometer to be operated at a nominal wavelength of 2.2 microns in the near-IR began in May 1987. Nearly all of the mechanical, electronic, and optical hardware of the Infrared Michelson Array (IRNMA) is currently in place and has been tested. Proof-of-concept has been demonstrated, and efforts are currently underway to improve the system operation to produce reliable, calibrated fringe visibilities.

In this second paper on IRMA (Infrared Michelson Array) the control and data acquisition system for the instrument are described. This paper will concentrate on a description of the computer-siderostat control system. Three personal computer based systems perform the high level control such as star catalog maintenance, current coordinate computations and siderostat/delay line position and rate computations, control of the analog to digital conversion, data storage and quick look data reduction. Low level control, such as generating the pulses that are sent to various drive micro-steppers are performed by seperate Motorola 68000 based microprocessor boards. Command words are passed between the control computers and the various microprocessor boards through a multichannel parallel-to-serial interface. These hardware systems and the software responsible for pointing and tracking with the siderostats, will be described in the text of this paper.

Two techniques we have used for determining mirror positions in multiple-aperture telescope arrays are described. Our design criteria is the maximization of the contiguous central core diameter of the optical transfer function for the telescope system. The first technique is applicable to arrays with a relatively small, O(10), number of apertures and is essentially an exhaustive search with a simple in-line test which allows the search space to be 'pruned' by an order of magnitude. In the second technique, arrays of a large number of apertures are designed using a fractal approach by recursively combining the results from several array patterns with fewer apertures. Both techniques are demonstrated for one and two dimensional designs and can be extended for higher dimensions if needed.

This paper considers the issues of lateral and longitudinal pupil geometry in ground-based telescope arrays, such as IOTA. In particular, it is considered whether or not pupil re-imaging is required before beam combination. By considering the paths of rays through the system, an expression is derived for the optical path errors in the combined wavefront as a function of array dimensions, telescope magnification factor, viewing angle, and field-of-view. By examining this expression for the two cases of pupil-plane and image-plane combination, operational limits can be found for any array. As a particular example, it is shown that for IOTA no pupil re-imaging optics will be needed.

Phase measurements made with the Mark III interferometer at Mt. Wilson using baselines up to 32 m show excellent agreement with the standard Kolmogorov theory, and give no evidence of an atmospheric outer scale smaller than 1 km. Thus, very long baseline interferometers (100's of m) can expect rms path length fluctuations to continue to grow nearly linearly with baseline length. With a wideband fringe tracker, atmospheric dispersion will cause significant reductions in fringe visibility for large instantaneous path length errors. A solution to this problem using an active dispersion tracker is presented. In addition, the problem of diffraction of the propagated beam, and the constraints it presents on the choice of beam diameter are discussed.

Optical interferometry has been shown to be a viable method for making high-precision astrometric measurements, with the capability for unprecedented accuracy in both narrow- and wide-angle regimes.' As astrometric resolution increases, however, the contribution of certain systematic internal errors of the instrument itself to the measured optical delay can become significant compared to the contributions due to angular position variations on the scale of 5-10 miliiarcseconds. In particular, the baseline of the interferometer can no longer be regarded as a fixed quantity at scales below about 1 micron. On the Mark III interferofneter at Mt. Wilson2, the major moving parts, aside from the optical delay lines, are the siderostats which mark the endpoints of the baseline. Non-ideal motions of the siderostat mirrors can cause changes in the instrument's baseline at the 1 micron level. To compensate for this motion, a prototype laser metrology system in the form of an optical tripod has been installed to measure changes in the instrument's baseline and supply corrections to the optical delay and delay offset data. This system has revealed a number of issues that are crucial to the development of future systems. In particular, future space-based systems3 will require laser metrology systems with accuracies of the order of 0.01-0. 1 nanometers, and the implications of the Mt. Wilson results on future systems are discussed, and current thinking on future designs is presented.

This paper describes the design and the capabilities of the Naval Research Laboratory Big Optical Array (BOA), an interferometric optical array for high-resolution imaging of stars, stellar systems, and other celestial objects. There are four important differences between the BOA design and the design of Mark III Optical Interferometer on Mount Wilson (California). These include a long passive delay line which will be used in BOA to do most of the delay compensation, so that the fast delay line will have a very short travel; the beam combination in BOA will be done in triplets, to allow measurement of closure phase; the same light will be used for both star and fringe tracking; and the fringe tracker will use several wavelength channels.

The Infrared-Optical Telescope Array (IOTA) is being developed by a consortium comprising Harvard University, the MIT Lincoln Laboratory, the Smithsonian Astrophysical Observatory, the University of Massachusetts at Amherst, and the University of Wyoming. The instrument is intended to generate high-resolution images of astronomical objects by bringing together beams from widely separated telescopes and combining them at a central location. The initial configuration will consist of two 0.45 m telescopes thay may range along an L-shaped track that will permit spacings in the 5 to 38 m range, at the Smithsonian's Fred L. Whipple Observatory on Mt. Hopkins. Initial tests of this configuration are expected to be conducted during the summer of 1991 and to yield both valuable engineering data and the first scientific results including diameters of stars and artificial earth satellites and a measure of the extent of some circumstellar shells. The engineering data will be applied to the refinement of IOTA, particularly to the second IOTA configuration, in which a third telescope will be added, making it possible to obtain phase closure information.

This paper considers in detail the questions of mechanical stability and thermal stability that were encountered in the course of the design of the Infrared Optical Telescope Array (IOTA) installed at the Smithsonian's Whipple Observatory on Mt. Hopkins. With respect to the mechanical stability, measurements of ground vibration showed that the ground is essentially benight at the IOTA site. To insure thermal stability, the telescope support points and the beam-relay path will be slightly below ground level and well-insulated. The telescopes will reside in insulated shelter-transporter structures which will be cooled during the day. The beam relay path will be in vacuum, so that seeing effects and longitudinal dispersion will be eliminated.

Pupil plane beam combination, spectral dispersion, detection, and fringe tracking are discussed for the IOTA interferometer. A new spectrometer design is presented in which the angular dispersion with respect to wavenumber is nearly constant. The dispersing element is a type of grism, a series combination of grating and prism, in which the constant parts of the dispersion add, but the slopes cancel. This grism is optimized for the display of channelled spectra. The dispersed fringes can be tracked by a matched-filter photon-counting correlator algorithm. This algorithm requires very few arithmetic operations per detected photon, making it well-suited for real-time fringe tracking. The algorithm is able to adapt to different stellar spectral types, intensity levels, and atmospheric time constants. The results of numerical experiments are reported.

Hybrid mapping techniques originally developed for radio aperture synthesis imaging are applied to the case of image reconstruction in optical interferometry. It is found that the use of redundant baselines with hybrid mapping can increase the fraction of phase information that is obtained in a realistic optical experiment to approach that obtained in a typical multistation radio experiment. However, this additional information does not greatly improve the quality of the resulting maps in cases with good coverage of the u-v plane. Thus, in constructing an experiment with redundant baselines, the adequacy of the u-v coverage should be considered along with any desire for redundancy. The use of phase differences between two colors in a dispersed fringe for hybrid mapping experiments are investigated and it is found that maps may be made of comparable quality to hybrid maps with phase closure.

To test the ability of the Infrared-Optical Telescope Array (IOTA) to 'map' complex objects, the IOTA project team plans to make 'ground-truth' observations of man-made objects in space. However, not all satellites and rocket bodies are suitable since there will be instrumental limitations on the brightness, angular size, angular speed, and the apparent structural complexity of the observed object. More than 200 objects with the requisite characteristics are identified. The GPS satellites, in particular, offer some interesting possibilities for instrumental testing and 'mapping'.

In a class of optical interferometers, two samples of the captured light are combined in the pupil plane, dispersed, and focused to form a channeled spectrum. Shifts of the resulting fringe pattern due, for example, to the rotation of the instrument or changes in the propagation medium, must be tracked and, in some cases, compensated for the instrument to perform properly. The performance of the fringe-tracking estimator sets the faint limit of the instrument for a given suite of disturbances to the optical path difference (OPD) through the two sides of the interferometer. A nearly optimal fringe tracker may be constructed as a Kalman filter that includes a correct representation of the statistical properties of the OPD time series and that processes the detected photons individually. The use of a suboptimal filter is likely to be necessary both because of the difficulty of properly representing the OPD statistics and because of the computational burden of the complete, nonlinear, photon-by-photon estimator running in real time. We discuss a portion of the study of fringe trackers that has been carried out at SAO for the IOTA (ground-based) and POINTS (space-based) interferometry projects. We also present the results obtained from a Kalman filter fringe tracker running on simulated data. When the simulated OPD is obtained by numerically solving a second order differential equation with a white Gaussian driving term, an augmented state extended Kalman filter substantially outperforms a similar filter with an unaugmented state.

A concept for a space-based interferometer dedicated to the detection of extrasolar earth-like planets is described. The interferometer is of the Fizeau configuration with an aperture composed of twelve 1.2 meter mirrors on a 20 meter ring. The subapertures are distributed to optimize the u-v plane coverage when the interferometer is rotated around its axis. Parent star cancellation is obtained by the combination of a coronagraph and a rotation shearing interferometer. The interferometer is supported by a chemically rigidized structure deployed by inflation. Due to the lack of side shield and the resulting limited pointing capability with respect to the sun, it is proposed to locate the instrument at the second Lagrangian point of the earth-sun system.

Speckle imaging is a powerful tool for high resolution astronomy. Its application to the core regions of globular clusters produces high resolution stellar maps of the bright stars, but is unable to image the faint stars which are most reliable dynamical indicators. The limits on resolving these faint, extended objects are physical, not algorithmic, and cannot be overcome using speckle. High resolution maps may be useful for resolving multicomponent stellar systems in the cluster centers.

This paper describes the preflight programs developed for two telescope missions, the U.S. Hubble Space Telescope and the European Hipparcos, in order to remove stars showing multiple nature from a set of proposed reference stars. It was found that, in a large unbiased sample of stars, 10 percent of stars at apparent magnitudes between 9 and 13 mag are resolved in the range 0.05-1.00 arcsec. The consequence is a significant increase in the difficulty of making observations with the space telescopes. In the case of Hubble telescope, this will add to the acquisition time of Guide stars by about 11 percent. In the case of Hipparcos, it will introduce small uncorrelated proper motion errors into a number of unidentified unresolved binary stars.

Algorithms for reconstruction of isoplanically blurred point source pairs are considerably simpler and faster than full-blown image reconstruction techniques. Traditional autocorrelation approaches suffer from a 180 degree ambiguity, however, and only yield order of magnitude estimates for brightness ratios. A new asymmetric algorithm is here presented: the "Directed Vector Autocorrelation" (DVA), which is a rapid alternative to vector autocorrelation. Together with the 'Fork algorithm", a directional filter for estimating brightness ratios, the DVA algorithm has been used to resolve ambiguous orbits and produce differential color photometry for several binary stars.

Diffraction-limited infrared images of three binary stars reconstructed using speckle imaging techniques are presented. Three methods of recovering the objects' Fourier amplitudes are compared, Fourier deconvolution, projection onto convex sets, and CLEAN. The objects' Fourier phases are retrieved via bispectral analysis. It is shown that the quality of the final reconstructed images is a function of the percentage of the total bispectral volume utilized in the reconstruction procedure.

The performance of a phase recovery algorithm developed for speckle data collected using a pupil-plane mask has been investigated for use at near-infrared wavelengths. The method, based on the radio-astronomical self-calibration technique, has been tested alongside a state-of-the-art implementation of the Knox-Thompson scheme using both simulated and real specklegrams. Results indicate that the new procedure is as effective as the Knox-Thompson based image reconstruction scheme and is applicable to a wide range of astrophysically interesting sources.

A program of high resolution infrared imaging at Palomar Observatory, is presented. The use of nonredundant masks both as an imaging technique and as a method of analyzing various aspects of the imaging system are investigated. In particular, the technique is applied to a bright star and binary system using a three-hole mask. The method is useful for understanding certain systematic biases in data, as well as in producing high quality images despite sparse UV coverage. The use of multi-r(o) apertures along with a large bandwidth does not significantly hamper image reconstruction, but provides significant extra coverage in the UV plane.

"Interspectroscopy" le a method of ol,taining the separated spectra of binary (or multiple) stars too close to be resolved by conventional techniques. The method is "passive" because, like speckle interferometry, the atmosphere provides a series of random phase variations, and no control system is used to maintain phase. Results in terms of spectral purity are given for several cases in both the pupil and image planes. It is shown that significant spectral separation can occur. We briefly discuss planned observations with a fiber-fed pulse-counting spectrograph at the 1.9-m DDO telescope.

The first application of stellar interferometry by Michelson was undertaken to make an angle measurement. Present and projected applications of the very precise angle measuring capabilities of modern interferometers include, in addition to stellar diameters: binary star orbits, parallaxes, positions, proper motions and the astrometric detection of extra-solar planets. A brief discussion of what has achieved and what can be achieved is given with emphasis on positional work leading to fundamental coordinate frames and motions within such frames.

For the first time, four spectroscopic binaries have been directly resolved with the Mark III Stellar Interferometer. Observations in 1988 and 1989 were analyzed, and visual orbits for four binaries have been determined. The semimajor axes for Beta Tri, Alpha Equ, Alpha And and Beta Ari are approximately 0.008 arcsec, 0.012 arcsec, 0.024 arcsec and 0.037 arcsec, respectively. The magnitude differences between two components are 0.5, 0.7, 1.8 and 2.6 mag, respectively. All of the orbital elements for Alpha And and Beta Ari were determined from interferometric data only, and agree well with spectroscopic observations. Predictions of relative position between the two components for these binaries are consistent with the measurements to less than 0.001 arcsec. Combined with data from spectroscopy, masses and distance for the double-lined spectroscopic binary Beta Ari are derived, and the results indicate that both components of Beta Ari agree well with the empirical mass-luminosity relation.

Reliable stellar angular diameters can now be determined using the Mark III Optical Interferometer located on Mt. Wilson, California. The Mark III is a Michelson Interferometer capable of measuring the interferometric fringe visibility for stars using interferometer baselines varying from 3 to 31.5 meters in length. Angular diameters measured with the Mark III Optical Interferometer are presented for 12 stars at wavelengths of 450 and 800 nm.

This paper is a presentation of the so-called COME-ON adaptive optics prototype system developed jointly by four European institutions. This system has been tested on the 1.52m telescope of the Observatoire de Haute Provence on October 12 to 23 and November 13 to 24, 1989. Diffraction-limited infrared imaging has been achieved during these first tests. The adaptive optics system consists of a 19 actuator deformable mirror and a Hartmann-Shack type wavefront sensor. In this instrument the wavefront sensing is performed at visible wavelengths while the correction is performed for near infrared imaging (1.2 to 5 rim). Specialized computers drive the deformable mirror and a tip-tilt mirror. The bandwidth of the servo-loop is 9 Hz at 0 dB point in open-loop. The results obtained with this instrument will be very useful for the design of the future adaptive optics system for the ESO Very Large Telescope (VLT).

Observational results obtained at a 1.52-m telescope in the near-infrared wavelength range of 1.6-5 micron with the use of an adaptive-optics prototype system are reviewed. The purpose of the experiment is to evaluate the correction performance of adaptive optics under actual turbulence conditions at an astronomical site. The observations demonstrate that the full gain in resolution can be obtained. A significant gain in sensitivity is observed, and it is shown that a considerable improvement in signal-to-noise ratio can be achieved due to the bright maximum of sharpened images in all cases where the signal-to-noise ratio is set by background or detector noise.

A telescope-mounted wavefront-folding interferometer designed to measure the phase fluctuations due to atmospheric seeing on a wide range of temporal and spatial scales is described. It makes use of photon-counting avalanche photodiodes to achieve high SNRs at high sampling rates. In addition to testing theories of the atmospheric turbulence responsible for the seeing, instrumental effects such as telescope vibrations can be quantified.

It is noted that by adjusting the phase of individual dither frequencies, the dynamics of beatings can be optimized, and normalization effects can be limited when dealing with wavefront distortions. A method for optimizing the beating dynamics by adjusting individual dither frequencies at the output of synthetizers is proposed. Two optimization methods are employed: one using the first derivative, and the other using no derivatives. The algorithm set derived indicates that in every case the local minimum point remains a limitation. It is pointed out that the chosen solutions which are closed to the global minimum point are often sufficient for the problem.

The purpose of the study is to establish the time transfer function of an adaptive-system loop with a CCD camera. Each component of the loop is replaced by its transfer function. The time lags due to the CCD matrix and calculations are determined. The effects of the sampling rate and the time lags on gain and bandwidth are studied by using the Nichols criterion defining the loop stability. It is concluded that the input should be phase perturbations expressed on the basis of the actuator correction shapes, while the output should be the voltage applied at the power amplifiers in order to obtain correction strokes.

A bimorph mirror seems to be low voltage, large strokes device which can be used as a correction mirror in an adaptive optics system for infrared applications. A few theoretical results are recalled and have been used to develop a numerical method to solve the displacements of a bimorph mirror supplied by a distribution of voltages. An example is given which involves seven electrodes; comparisons with theoretical and other numerical results are achieved.

The servo-loop components and control algorithm of the COME-ON project involved in imaging in the near-infrared range with a 4-m telescope are described. Attention is focused on the wavelength sensor, wavefront computer, control computer, high-voltage amplifiers, and deformable and tip-tilt mirrors. An interaction matrix and its modal decomposition are discussed, along with modal filtering and tip-tilt control. It is pointed out that the comparison between theoretical and experimental eigenmodes of the interaction matrix demonstrates the spatial modelization validity of the system.

Recently, an iterative deconvolution algorithm has been priori object constraints on speckle interferometric data. This paper demonstrates its application to both simulated and real two-dimensional infrared speckle data and compares its performance to that of an iterative transform algorithm. This iterative deconvolution algorithm differs from the iterative transform algorithm in that the object's Fourier modulus, as well as the Fourier phases, is also constrained by the a priori information.

Integral representations of the average bispectrum, the bispectrum variance, and the bispectrum covariance are developed for onedimensional infrared speckle data. The integrals are evaluated by Monte Carlo integration for a particular case. The results are compared to the corresponding sample quantities computed from simulated data.

This paper describes a simulator we have developed to model speckle imaging and resconstruction of astronomical objects. The simulator was designed as a tool in the development of new signal and image processing techniques for our high resolution imaging research. It has been found to accurately replicate the speckle imaging process, and can be used to predict experimental results under various environmental conditions. The simulator is described in detail, including the modeling of atmospheric turbulence effects, the generation of speckle images, and the simulation of the telescope and image detection processes.

Optimization techniques are applied to the bispectrum. The Levenberg-Marquardt algorithm for nonlinear least squares is used to determine the parameters of binary stars. The conjugate gradient and conjugate direction algorithms are used to estimate the full set of object phases and moduli. Simulated and observed data are utilized in each optimization.

The computer system described in this paper is designed to capture event data from a photon-counting speckle camera at photon event rates of up to 1 MHz continuously. The display and quicklook computer uses several single board computers (SBC's) to display the photon events in real-time, calculate the centroid of the data for autoguiding of the telescope, and calculate the autocorrelation function. The system is based on the VMEbus architecture. The SBC's operate under the VxWorks real-time operating system. A Sun workstation is used for code development. the SBC's are mostly selected for speed since the computational requirements are very high. Eventually a Sun workstation for near-real-time image processing and image reconstruction will be used to receive quicklook data from the control computer.

Hardware and software improvements to the IR speckle camera are reported. The observing experience obtained during the first year of operation allows a preliminary discussion of exposure times, limiting magnitudes, observing strategies and problems, duty-cycle, data handling, and real-time and off-line processing. The results with this system have also helped to define directions for future developments in high resolution IR imaging.

A four-stage Proxitronic® image intensifier system with digital video readout provides high detective quantum efficiency photon-counting imaging. The objective is to demonstrate a prototype system for further evaluation of image intensifier capabilities in speckle imaging applications, and to demonstrate an observing system opumized for use in diffraction- and statistics-limited imaging with accurate noise bias calibration at the lowest possible photon flux with the highest possible photon-counting efficiency. The phosphor image is Kowa® transfer-lens coupled to a Pu!nix® CCD video camera at 2: 1 reduction to preserve high spatial frequency response. The data are digitized at the telescope and fiber-optic coupled synchronously to a digital video processor which can operate with non-interlaced full-frame subtraction in linear grey-scale analog mode as well as in photon-centroiding event-counting mode. A MathCop2® 18MHz floating point math coprocessor provides quick-look capability for near-realtime inspection of image spectra or Knox-Thompson cross-spectra, using FFF methods for grey-scale data or vector autocorrelation for event data.

Results will be presented from a horizontal path imaging experiment in which we used a 0.5 meter telescope focused on a target located at a range of 1 .2 km. The targets included various extended objects from simple binary letters to extended representations of satellites with grey scale and size variations. Imaging at a center wavelength of 0.7 microns, we found an atmospheric degradation factor of Dir0 = 17, on average. We used a slow read-rate bare CCD detector and thus had to effectively deal with additive noise in the speckle measurements. Our image reconstruction algorithms are based on the use of the complex bispectrum and we have demonstrated diffraction-limited imaging down to light levels approaching a few photons per speckle per resolution area. We have paid careful attention to the effects of additive noise on the reconstruction process and shown that they can be adequately overcome.

With the recent availability of low-loss single-mode fluoride glass fibers, it becomes possible to consider interferometric beam recombination with fibers at infrared wavelengths. After a brief description of their present performances, specific aspects of the use of single-mode fibers in the infrared are studied here. In an astronomical interferometer thermal emission from the fibers will limit the sensitivity if they are not cooled. It is interesting to use the fibers at large X/? ratios: the beam is wider inside the fiber, most of the energy being routed by the cladding, and the coupling to starlight is eased because it can be done at slower f-ratios. This is true as long as X/X<2; beyond, evanescent field leaks through the fiber coating at the bends become too strong and prevent any practical use of the fiber. Far field beam profiles have been measured in the K and L bands. A project for a laboratory demonstration interferometer is presented: fringes were obtained in the visible and an extension to longer wavelengths is under way.

We present the first experimental results of recovery of the 2-D spatial coherence modulus and phase using fourth order correlation interferometry (FOCI). The technique, a generalization of intensity interferometry and laser speckle correlography, measures the correlation (I(x)A*(x+A)A(x+A+c)), where 1(x) and A*(x+E)A(x+A+e) are measured with separate, unconnected apertures. The result of the measurement is t*(A)i(+c), a cross-product of the spatial coherence factor that is analogous to the Knox-Thompson cross-spectrum in astronomical speckle imaging. The technique has applicability when the coherence function must be measured over distances A which are larger than the diameter D >> XE of the largest diffraction limited optic or amplitude interferometer and the field is a circular complex Gaussian random variable. An experiment is described which makes the correlation measurement on a series of laser speckle patterns. 1(x) and A*(x+A)A(x+A+e) are obtained from the focal spots of an array of lenslets. The measured coherence function is shown to agree with the expected value. The intensity image, given by the Fourier transform of the coherence function, is calculated and gives good agreement with the experiment target. The usefulness of the technique is discussed.

Speckle interferometric imaging techniques using image power spectra to obtain calibrated image amplitudes are contaminated by noise biases inherent to the detectors themselves. This noise bias contribution is non-linear for detectors in a conventional image amplitude detection mode (used for bright objects--over 10^5 photons/sec--such as SN1987A, for example). This paper presents our successes in modelling this non-linear bias contribution with iterative minimization techniques.

A multiple aperture optical instrument capable of creating a high resolution image and spectral decomposition of a scene is described and analyzed. The instrument operates under wideband illumination, uses small apertures that are widely spaced, and incorporates electronic and digital processing to obtain high quality images and to perform spectral analysis.

We have used a laser interferometer to measure the relative optical pathlength variations from the focal plane to the six individual secondary telescopes of the Multiple Mirror Telescope (MMT). These pathlengths vary as func- Lions of elevation due to variable gravitational loading, drive acceleration and velocity, temperature variations, and wind loading. Vibrations induced by wind loading and telescope drives, including building vibrations transmitted to the mount, cause high-frequency variations in the otherwise slowly varying optical pathlengths. This experiment was designed to evaluate the effects of these high-frequency perturbations on optical interferometry at various tracking rates, including those relevant to Earth Satellite observation. We find effects which can strongly affect the contrast of interferograms.

A method for imaging extended objects using adaptive optics technique is proposed. The wavefront distortion is corrected by using the Fourier transforms of two image intensities obtained with and without a Gaussian amplitude filter at the entrance pupil plane of the imaging system. A one dimensional computer simulation is presented, which shows the effectiveness of the method.

A laser interferometer has been used to measure the relative optical pathlength variations from the focal plane to the six individual secondary telescopes of the Multiple Mirror Telescope (MMT). These pathlengths vary as functions of elevation due to variable gravitational loading, drive acceleration and velocity, temperature variations, and wind loading. Vibrations induced by wind loading and telescope drives, including building vibrations transmitted to the mount, cause high-frequency variations in the otherwise slowly varying optical pathlengths. This experiment was designed to evaluate the effects of these high-frequency perturbations on optical interferometry at various tracking rates, including those relevant to earth satellite observation. Effects are found which can strongly affect the contrast of interferograms.

An infrared camera has been developed for interferometric imaging on large telescopes. Observations obtained with the 6.86 m aperture of the cophased Multiple Mirror Telescope (MMT) demonstrate the ease with which future 8 m telescopes can achieve diffraction-limited performance from 1 to 5.5 micron. With the MMT, the infrared camera has imaged astronomical sources at 3.5 micron with a diffraction-limited resolution of 0.10 arcsec. Adjustable exposures as short as 4 msec are obtained at a maximum rate of 10 Hz to freeze atmospheric turbulence.

This paper describes the designs of the delay lines constructed for the Mark III Optical Interferometer (M3OI) currently operating on Mt. Wilson and for the Infrared Optical Telescope Array (IOTA) to be constructed for Mt. Hopkins. Special attention is given to experience with M3OI which was used to improve the design of the IOTA system. Schematics are presented for the IOTA instrument layout, the IOTA coarse delay line, and the IOTA long delay line.

The operation of the GI2T two-telescope inteferometer, a precursor to the Optical Very Large Array (OVLA), is described. The GI2T, consisting of two 1.5-meter telescopes, is beginning to provide high-resolution data on stellar envelopes. It is also used to test components for the 27-telescope OVLA. Prototypes of the metrology system, the compact telescopes, and a walking robot are under construction. These developments are preparatory to the construction of the OVLA and its moon-based version with three to 33 telescopes. Following tests of tracking and vibration performance, the prototype 1.5-m telescope will operate with the GI2T.

Pros and cons of Michelson and Fizeau type interferometer configurations are compared for systems in space having baselines in the 5 to 100 meter range, when a large field of view is requested for off-axis tracking. The size of the coherent field-of-view is larger in a Fizeau type interferometer but the tolerances required by wide field operation are easier to achieve in the Michelson type. Furthermore, aberration compensation, which calls for more than three optical elements per beam to assure that the beams overlap in the Fizeau field of view, makes it necessary to split the on and off-axis fields before the f/250 final combination plane in both cases. Two categories of configurations are defined: for baselines B less than 10 meters, a Fizeau type is preferable; conversely a Michelson type should be adopted when B is greater than 10 meters. After summarizing the difficulties to overcome in each type, a solution for each type in space environment is proposed.

We report preliminary results on fringe pattern acquisition and stabilization as performed on a Mach- Zehnder set up representative of Stellar Interferometry needs. The system algorithm is based on "white light" fringe tracking controlled from a reference interferometer synchronous detection (central fringe locked interferometry). This servo-system drives a simple two-stages delay line for real time compensation of the optical path delays either due to atmospheric or instrumental errors. Contrasts (nonoptimized) of 70 % on the stabilized fringe pattern were measured, strongly encouraging to pursue the development of the technique.

We briefly recall the characteristics and particularities of the optical recombination of the Solar Ultraviolet Network (SUN) experiment which consists in a non-redundant << compact >> array of 4 telescopes of 20 cm diameter aligned on a 2 meters linear baseline. We proceed to the evaluation of the geometrical and optical aberrations of the recombination outlining the difficulties inherent to interferometric systems (in particular, and forgotten up to now, the << cophasing condition >>).We conclude on the advantages linked to a compact system like SUN.

Knox-Thompson Speckle Imaging is studied when partial adaptive optic compensation is employed. A generalization of previous analysis allows treatment of the compensation, finite exposure time, anisoplanatism and non-zero spectral bandwidth within the extended Huygens-Fresnel framework. The compensation process is assumed statistically stationary and accounts for wavefront sensor photon noise. The effect of finite exposure time is treated assuming Taylor's hypothesis and the Bufton wind model. An innovation is a perturbation treatment of fmite spectral bandwidth effects accurate to Significant improvement results from partial compensation when 10 degrees of freedom can be corrected with < lrad2of phase error due to shot noise. The compensation also allows significantly greater spectral bandwidth for the speckle images relative to the uncompensated case. Strategies optimizing over compensation scale, spectral bandwidth, and exposure time are chosen to optimize the SNR for a fixed observation time. The results of the analysis are compared to a computer simulation with a wave-optics propagation code.

In this note, I examine the properties of various pupil plane schemes of beam-recombination for a non- monolithic interferometric array of telescopes. I first recall the basic features of the Modulus Transfer Function in the pupil plane, and discuss the phase estimation through phase closure techniques. I aIso address the questions of the interferometric field of view and of the combination of pupils of unequaJ sizes. Finally, I propose some recombination schemes adapted to the various wavelength domains.