Aircraft-mounted inertial navigation systems (INS) use stellar observations to improve accuracy of position and attitude. The INS provides a stable platform on which is mounted a two-axis gimballed telescope, which may be pointed to any of 57 stars of magnitude 2.5 or brighter. Only one target star appears at a time on the 6x6 arc-min field of view (FOV), with speeds up to 11 arc-min/sec. Previous implementations used optomechanical techniques to separate a star from a sky of up to 3000 fL luminance, allowing estimation to 5 arc-sec accuracy in the daytime and on the ground. This paper describes a more reliable implementation using a charge-coupled device (CCD) and high-speed microprocessor. Out of 385x288 pixels in the image section, only 30x30 are digitized each frame time of 10 ms. To reduce background shot noise, the charge accumulation duty cycle is varied so that pixel wells fill each frame regardless of sky brightness. Frame times are kept short to reduce dark noise and track fast-moving stars. During each frame time the microprocessor corrects each pixel for responsivity and background charge, averages over several frames, and forms a multipixel window. It then either searches an empty field for a star or tracks and interpolates to estimate the centroid of one already found. The sensor requires 52 in.2 PC board space and 4 W dc power. Ground tests on the first pre-production unit exhibit superior performance to its predecessor, tracking 2.5 magnitude stars in as much as 4000 fL urban skies.

A mobile optical test facility (MOTF) has been constructed for field testing optical pattern recognition systems. Currently a VanderLugt correlator is used to detect and track targets of interest. A nonlinearity in position of the correlation signal with respect to the location of the target has been detected and explained in terms of fundamental optics. This nonlinearity is present in any VanderLugt system and ultimately limits the tracking accuracy. A miniature optical correlator, mounted on gimbals, has demonstrated the ability to use correlation signal motion as a feedback to maintain accurate pointing of the sensor. Noise rejection and false correlations are characteristically not a problem for VanderLugt optical correlators, and targets occupying about one percent of the field of view have produced correlation signals well above background levels. A different correlator architecture, commonly referred to as the joint transform geometry, is being investigated in the laboratory while field testing of the VanderLugt correlators progresses. This arrangement does not require matched filters made in advance, but may not provide the required signal to noise when the target is surrounded by clutter. The ultimate goal of this research is to test the direction, pointing and tracking abilities of several optical correlators, all located in the MOTF , against a common target in the presence of background clutter.

The TPD Off-Axis Star-Tracker (TOAST) aims at achieving a high accuracy of 1.5 arcsec in a 5.4°x8.1° field of view, combined with random pointing capability. This requires a guide-star limit of msilicon = 6.5, giving a 99% probability of finding at least two guide-stars in the field of view, thus permitting 3-axes attitude reconstitution. To enhance the usefulness of the sensor three modes of operation are provided, enabling respectively a survey of the starfield, guide-star acquisition and fine tracking. The sensor is based on a 576 x 385 elements CCD detector. A microprocessor electronics enables an efficient organization of the detector readout and starimage processing, rendering subelement starposition accuracies. The detector is cooled thermo-electrically. A functionally representative breadboard model was built and subjected to performance tests. Test results show that the measured performances are not far from the design goals.

Boresighting involves precisely transferring angular coordinates from one point to another. This frequently arises in correlating directional information between systems such as optical telescopes or high-gain microwave antennas that are physically separated by a truss or support. Generally, the separating truss is not rigid to the accuracies required for precise alignment because of thermal, aerodynamic, and vibrational distortions typical of many practical environments. This paper presents a relatively simple concept for optically boresighting between any two points, in three degrees of angular freedom, and independent of the rigid body. Analyses and laboratory investigations indicate that sub-microradian accuracies, data rates exceeding 100 Hz, and dynamic ranges of approximately milliradians are feasible between points separated by over 10 meters.

The Space Infrared Telescope Facility (SIRTF), a NASA "Great Observatory" to be launched in the late 1990s, is a superfluid helium-cooled one meter class IR telescope with a sophisticated chopping facility provided by a dynamic two-axis tilt control capability of the secondary mirror. This paper describes the pointing performance analysis of the Prototype Secondary Mirror Assembly (PSMA) design. The two-axis PSMA tilt control system employs four linear actuators and four pairs of position eddy current sensors, and a reaction mass system to isolate the servo loops from the structural modes. The actuators and sensors operate in the 4 K environment of the secondary mirror assembly, while the control electronics reside in the "warm" electronics box outside the dewar. The PSMA design meets stringent pointing performance requirements over a range of discrete chop amplitudes and frequencies, for three different dynamic chop modes. The analysis of the pointing performance utilizes a detailed dynamic model of the PSMA in time and frequency domains, including the discretization and quantization effects in the servo-controllers, the sensor noise, and the structural modes. The servo stability issues are also addressed.

The active optical elements of the Goddard High Resolution Spectrograph (GHRS) consist of five plane diffraction gratings, one Echelle and four non-dispersive mirrors. The instrument has eleven modes for ultraviolet spectroscopy and target acquisition. The gratings and mirrors are mounted on a precision mechanism which rotates to position the desired element in the optical path. Fine adjustments select the precise wavelength region of interest and control a variety of scanning options. Aside from a simple two position shutter this is the only moving part involved in normal science observations. The requirements for precise image location, high spectral resolution and efficient autonomous operations place stringent demands on its performance. The carrousel is rotated by a brushless DC torque motor, which produces a slew rate of six degrees per second. The 16 bit position encoder provides a granularity of approximately 20 arc seconds per encoder step. The active electronic control system maintains pointing stability with an RMS jitter of 0.13 arc seconds, which corresponds to 0.033 pixels of image motion at the detector. The repeatability is such that the image returns to within ±0.10 pixels of the same position over 90% of the time. After rotating to a new position the carrousel settles to within the jitter within 3 seconds. This paper describes the mechanical and electronic design, the hardware logic and flight software algorithms which control its operation. Data from pre-flight test programs illustrate the performance of the carrousel.

This tutorial paper discusses the performance requirements and design parameters for Electro-Optical Tracking Systems. Descriptions of tracking systems for range instrumentation are discussed. The relationships between the performance requirements, target characteristics and the system optical sensors, tracking algorithms, pedestal dynamics, servo control loops, and ancillary equipment are examined. The design considerations related to the operator interfaces are also discussed.

Quantitative evaluation of anti-ship seeker design parameters can best be accomplished with a generic hardware seeker with variable parameters that can be operated in a real world target environment. This paper describes an infrared seeker system that can be modified to incorporate different platform, detector, optics and signal processing characteristics. This guidance simulation equipment is operated in a captive mode on a fixed wing aircraft and is used to evaluate the simulated seeker acquisition, track and guidance performance as a function of the target range and signature characteristics. Real time operator monitoring of the performance is accomplished through seeker video images which are recorded for subsequent laboratory analysis. Modifications to the guidance unit target detection and track algorithms can be accomplished between data runs by software changes to a dedicated image processing computer. Descriptions of the sensor, platform and image processor are presented. The range of parameters that can be simulated are discussed and typical data from operational tests are presented.

Image motion compensation (IMC) can be used to augment conventional line-of-sight (LOS) stabilization systems to improve stabilization performance. The tradeoffs involved in deciding when to use IMC and the expected performance improvements are reviewed. Several IMC concepts are outlined including the special case of a gimbaled optical element with a one-to-one scale factor between mechanical and image motion. The advantage of this IMC stabilization approach over other IMC techniques is that performance is not limited by the bandwidth of the sensor or actuator. Performance, however, is limited by optical scale factor accuracies. Precision measurement of the gimbal motion is not required to achieve LOS stabilization. The "restored mass stabilization" approach offers performance improvements without the high cost of precision inertial components.

Development of precision, unmanned mobile tracking and pointing systems subjected to rough vehicular disturbances requires the use of robust, wideband, optimal-adaptive-intelligent motion stabilization control systems. Practical implementation of such theory is impeded by hardware induced constraints such as mechanical resonance, noise, and plant uncertainties. Several techniques for overcoming these hardware constraints have been demonstrated in Lockheed's long term Controls R&D Program for developing advanced mobile tracking and pointing systems. This paper presents the findings of a recent study to determine the applicability of the Linear-Quadratic-Gaussian/Loop-Transfer-Recovery (LQG/LTR) design methodology to the development of wideband gyro stabilization control of mobile electro-optical sight systems. In the presence of high order plant dynamics and uncertainties, LQ control methods are normally confined to low bandwidth loops. The ability of the LQG/LTR methodology to generate robust, wideband, optimal motion stabilization control of plants containing high frequency dynamics and random base motion disturbances is described. An overview of the LQG/LTR design procedure is presented for background knowledge. Use of the LQG/LTR design technique to modify control of an existing gyro stabilized EO sight system is discussed and study findings are reviewed.

This presentation summarizes the analytical work that has been preformed to date by the Ballistic Research Laboratory in evaluating proposed conceptual fire control systems which utilize modern control theory to enhance: Stationary Firer/ Maneuvering Target, and Fire-On-The-Move Performance. Both conventional man-in-the-loop and automated tracking sytems were used in conjunction with the conceptual fire control systems. Much of work was performed for the US Army Research Development and Engineering Center, ARDEC, as part of that Agency's Very High Speed Integrated Circuits (VHSIC) Program. The presentation begins with a brief overview of the different types of control systems currently being employed in US and foreign production and testbed armored vehicles. The following systems are treated in a generic manner: Driven Reticle, Gun Director, Free Floating Gun Director. The presentation contiunes with an evaluation of each of the above fire control systems' stationary firer performance against maneuvering targets. The fire-on-the-move anlysis follows, with a discussion of the digital optimal controller that was used in place of conventional classical analog controller. To date only the manual tracking system has been evaluated. However, an automatic tracking system using a FUR tracker will he evaluated in the near future.

An expression for the probability density function for a doubly stochastic Poisson process driven by illumination beam jitter was derived. The formula was used to calculate the effect of jitter on the probability of detecting a point target in presence of both illumination beam jitter and additive background noise. The illumination beam is assumed to have a Gaussian profile with standard deviation σB. The jitter distribution is also Gaussian with standard deviation σj. It was found that for the quantity σj/σB, called the jitter ratio, exceeding 0.5 that the effect of jitter becomes significant.

This paper discusses the results of a flight test conducted by Ball Aerospace Systems Group (BASG) to evaluate, for reconnaissance and surveillance purposes, a forward oblique scanning sensor (FOSS) and an infrared line scanner (IRLS ). The paper discusses the sensor analysis, hardware/hardware installations, test objectives, procedures, and test results. Both a stabilized-gimballed framing sensor and an IRLS were flown during the spring of 1988 over a variety of civilian and military targets. The FOSS was microprocessor-controlled, enabling the framing imaging sensor to cover a wide field of regard (FOR) with a high-resolution narrow field-of-view (NFOV) sensor without gaps in the ground coverage. The IRLS had a 180-deg FOV, moderate resolution, and was flown simultaneously with the FOSS. Results of ground coverage modeling for the FOSS and IRLS are also presented. The comparison results are summarized and parameters such as video presentation format, off- and on-axis performance, terrain effects, ground processing techniques, and V/H are discussed.

As part of the Strategic Defense Initiative (SDI) program, Brookhaven National Laboratory (BNL) has constructed a Neutral Beam Test Facility (NBTF). This facility uses surplus capacity from the 200 MeV H- linac injector for the Alternating Gradient Synchrotron (AGS). An H- beam can be transported into the experimental area via a beam transport tunnel containing magnetic optics and a beam spectrometer. A reference target system, situated at the end of a 100-meter vacuum flight path extending from the experimental area confirms the pointing direction of the particle beam. This reference target system can also provide pointing direction information for laser beams. The facility is normally operated with a particle beam energy of 200 MeV and a beam current of approximately 0.5 mA (4.5 pps, 450 μsec pulse length) transported to the experimental area. Operation at lower beam energy is possible but at reduced beam current. A low divergence laser beam may be folded into the particle beam path as part of an ATP-related experiment. The facility is now in operation, and an ATP experiment is scheduled to run during May 1989.

High-fidelity simulation of automatic video trackers is needed to accurately train personnel in the use and operation of target track point update, tracker mode selection, and break-lock conditions. This paper describes a tracker simulator consisting of a microprogrammable two-card set, replicating the algorithms of four operational trackers, that has been developed for this purpose. The tracker simulator operates on images generated by a Computer-Generated Synthesized Imagery (CGSI) system in both the electro-optical and infrared spectrums. Servo stability and track loop delay characteristics are presented.

The use of an infrared imager in an air-to-ground tracking and guidance application is considered. The application requires tracking certain small distinct hot or cold spots in the presence of other potentially confusing objects. The target spot must be distinguished from the surrounding clutter. A tracking method based on pattern recognition techniques is used. Several approach sequences of IR imagery taken during captive flight were recorded digitally and used for this development. For each approach sequence, the tracker was initialized on a target spot. During each image frame, several feature measurements were made of the target spot. Confusing objects in the tracker search area were also identified, and the same features were measured for each detected clutter object. The target and nontarget feature measurements from all frames of all sequences were used to formulate a two-class pattern recognition problem. The classes were found to be well separated by both linear and quadratic discriminants. When the derived discriminants were used to classify the target spot from nontarget objects, high correct classification probabilities and small false alarm and miss probabilities were obtained. This resulted in more accurate tracking performance, even in scenes containing a high density of confusing clutter objects.

We are developing a wide-field-of-view (WFOV) telescope system for military, commercial and scientific purposes. The system will have smart focal plane processors to recognize and track moving objects. This paper describes candidate algorithms intended for these processors which are currently implemented on a commercial image processing system interfaced to our prototype WFOV telescope.

Estimation of the angular position of point objects is considered for an active tracking system consisting of an agile scanning transmitter and a photon-bucket receiver. Bounds on tracking accuracy for an arbitrary scanning pattern and parameters of an optimal pattern in precision tracking are derived. Examples of linear and nonlinear algorithms for position estimation, with accuracies approaching theoretical bounds, are presented.

The problem of tracking N targets, with correlation in both measurement and maneuver noise sources, is solved by transforming to a coordinate frame in which the N targets are decoupled. For identical targets, the decoupling is shown to coincide with a transformation to a set of nested center-of-mass coordinates. An example with a white-noise target acceleration model is worked out in detail in order to demonstrate the significant improvement in both absolute and differential tracking accuracies that is achieved by properly exploiting the correlation between targets.

Detection and position estimation of targets in the presence of strong scattering background clutter with an optical Doppler Autodyning Tracking (DAT) technique is proposed for the first time. The DAT technique is based on measuring relative Doppler shifts introduced by different velocities of target and background scatterers with a direct detection method.

A novel architecture for a ladar transceiver is proposed to achieve high-resolution imaging and precision tracking of complex targets. This architecture combines a conventional imaging receiver (telescope) with a noncoherent sparse array of photon buckets into one system. The diffraction limited resolution of the array exceeds that of the telescope. This ladar architecture is designed to use a new technique, CRITIC, which is based on combining the relevant information obtained simultaneously with the telescope and the array. In particular, the CRITIC technique allows one to overcome :several drawbacks of conventional monostatic ladars and opens new prospects for both the ground- and space-based ladars.

The use of multiresolution methods for the detection and segmentation of objects in images has been widely examined in the literature. The approaches used have largely been concerned with computationally expensive iterative pyramid linking procedures, and it is only recently that less costly procedures have been investigated using, for instance, top-down tree-growing methods which are computationally more efficient than earlier techniques, and which provide comparable detection and segmentation performance. This paper is concerned with the use of computationally efficient hierarchical techniques for object detection and segmentation, and describes several such algorithms, both new and previously published, which exploit the pyramid structure using vertical interactions between levels. The algorithms use mainly top-down approaches to achieve good performance at a lower cost relative to earlier techniques. Specific problems associated with automatic initialisation and start node selection are also addressed. The algorithms are discussed and their performance on both synthetic images and real infra red images is compared in terms of segmentation quality and computational cost. Results using the earlier iterative linking procedures are also presented, and are compared with the present algorithms in terms of cost and performance.

This paper describes the design and fabrication of a self contained satellite beacon tracking system for optical communications. This device is installed on a tracking mount and uses high bandwidth steering mirrors to effectively cancel atmospheric turbulence and mount jitter. The steering mirrors are integrated with a 20-cm, unobscured telescope used for receiving the beacon and for transmitting a modulated communications uplink. Satellite range and velocity calculated for its ephemeris are used to offset the transmitted beam from the beacon by as much as 50 microradians to compensate for optical transit time.

This paper is concerned with the architecture and functional analysis of the Pointing Acquisition and Tracking (PAT) subsystem of the European Silex program (Semiconductor laser Intersatellite Link Experiment). The Silex scenario is composed of two optical terminals mounted on low earth orbit, and geosynchronous orbit spacecrafts. High rate communication between the two terminals requires very accurate tracking performances over a high bandwidth taking into account satellite vibrations as well as PAT electromechanical or optical noise. In a first part, a detailed presentation of the retained functional architecture in terms of control laws and strategies, steering algorithms and related signal processing is given, while a second part presents the design choices retained for the PAT equipments. Emphasis will be given on the different operating modes strategy and switching required by the different phases of the PAT mission : acquisition, tracking and pointing. Finally, a performance analysis is carried out that concerns the rejection capacity with respect to vibrations generated by the host spacecraft, as well as optical sensors noise and mechanism error effects. Spacecraft vibrations profile is derived from in flight measurements on SPOT1 satellite and on ground measurements on SPOT2. Performance analysis will be illustrated by time and frequency simulations based on equipment modelling, and actual performances measurements derived from current associated technological programs.

The SILEX project is an experimental communication system aimed at demonstrating, in orbit, the feasibility of intersatellite optical communications using semiconductor lasers. As part of this project we have developed a precision mechanism to point the transmitted beam ahead of the current receiving satellite position. This is necessary due to the relative motion of the satellites, the narrow beam and the finite velocity of light. The design and construction of a prototype of this device is discussed along with measurements of performance. The technique as described can be used in many applications requiring precision beam stearing or rotation control.

Laser ranging to lunar surface retro-reflectors has provided a valuable source of data for the investigation of the lunar orbit, the earth's orientation in space, general relativity, and many other aspects of solar system dynamics.1,2,3 Due to the moon's relatively large distance, and the requirement to keep the beam from the relatively low-power laser very narrow, acquiring this data type requires very accurate telescope pointing and tracking capabilities. The nominal requirement is to keep the telescope on target within approximately 1 arc-sec of its predicted location for at least several minutes. An expansion of this requirement and the methods we have used to accomplish this high precision pointing are presented. Difficulties encountered and a few future goals of automating the McDonald station are also discussed.

NASA's shuttle-borne Broad Band X-Ray Telescope (BBXRT) consists of two glancing incidence imaging mirror assemblies mounted on an optical bench which is bolted to the primary structure of the instrument. The X-ray detectors are located in the focal plane of the mirror assemblies approximately 3.5 meters away. It is desirable to monitor the relative alignment of these components throughout ground testing, and to determine the magnitude of launch or thermally induced perturbations to the alignment during flight. The Displacement Monitor System (DMS) was designed to accomplish this task. This paper describes the design of the DMS, the development and optimization of the DMS cablibration facility, and the characterization of the system. The characterization of the DMS includes environmental qualification, displacement vs output calibration over the operating temperature range, a detailed error analysis, and the generation of a calibration polynomial which utilizes DMS detector output and thermocouple data to optimize system performance. The DMS accuracy exceeded the requirements of a 15 arc second limit of error, and passed the stringent environmental tests. As such, the DMS is one of the first flight qualified displacement monitor systems with this accuracy to be flown in space.

The Goddard High Resolution Spectrograph (GHRS) has two square entrance apertures, one two arc seconds and the other one quarter arc second on a side. Light from the astronomical target must be focused onto one of them prior to beginning a spectroscopic observation. The acquisition process is initiated with the target in the large aperture, followed by a slew to center the target in the small aperture if desired. The GHRS software supports both observer interactive and autonomous on-board modes of target acquisition. Software exists which executes a search pattern by commanding slews of the spacecraft, makes high resolution maps of the field of view of either aperture, recognizes the presence of the desired object, measures its precise location within the aperture, computes and commands a centering maneuver, and measures the flux of light at the conclusion. For most of these functions there are several options or algorithms, providing the flexibility to support a wide range of brightness and spectral content, extended or point sources, isolated or crowded fields, stationary or moving targets. Most acquisitions will be performed in undispersed ultraviolet light, although the capability to image the target in the light of an individual emission line is under development.

The fine pointing information required for the extraordinary pointing stability of the Hubble Space Telescope (HST) is provided by the fine guidance sensor (FGS), which is capable of measuring extremely small pointing errors, of the order of 3 milliarcseconds. The FGS provides this fine pointing information over a relatively large, continuous range of angles of guide stars relative to the optical axis of the telescope. The FGS accomplishes this task by a judicious design involving the following: a two-axis high precision star selector servo system; an optical deflection system that provides high reduction from shaft angle to optical angle; a unique interferometer using Koester prisms and extremely sensitive photomultiplier tubes that enable the FGS to lock onto guide stars whose magnitude may be any value over a wide range of brightness (including 14.5 Mv and fainter); and the use of the large aperture of the telescope itself to advantage. This paper describes the FGS control systems involved in achieving this performance, supported by test data.

The line of sight for optical beams from a spacecraft for pointing and tracking can be controlled in several ways. The whole optical system can be pointed to the object of interest, such as for the Hubble Space Telescope. The optical elements, such as a telescope, can be gimbal mounted and maneuvered with respect to an independently controlled spacecraft base, such as the European Space Agency Instrument Support Assembly. A third method maneuvers mirrors so as to direct the line of sight through spacecraft mounted telescope optics. This paper addresses the last alternative, Extended response agility can be accommodated with multiple mirrors. The field of view can be established with a relatively slow gimbal mounted mirror with a considerable angular range to be augmented in the optical path with a relatively fast response angular mirror with a limited angular range. The use of the error signal from the slow mirror to command the fast mirror extends the effective bandwidth of the system.

Estimation of positions, amplitudes, and the relative phase of double star targets is considered for an imaging system when the image pattern is registered by an incoherent array of photodetectors. An analytical framework is developed to establish the fundamental limits on the resolution of an imaging system and to obtain a quantitative understanding of the dependence of the accuracy of estimation upon the separation between targets and on the number of photons available. The additional number of signal photons required to obtain a desired accuracy for any parameter when the number of unknown parameters increases is computed and discussed for selected cases. Targets representing sources which can be either mutually coherent or incoherent are considered.

A Boiler & Chivens astronomical mount has been modified and upgraded to enable precision, tracking of near-earth objects. A. two-step process is employed: 1) ephemeris-based pointing to bring the object in or near the telescope field-of-view, and 2) automatic, video tracking (AVT) to "grab" and center the object. Tracking rates, in excess of 3°/s in each coordinate have been achieved. When operated in the AVT mode, observed pointing errors, even at the above rates, rarely exceed 1 pixel (2 - 5 arcsec).