Most deployed , surface-to-surface, antitank missile systems do not use passive homing. These systems require an operator for target tracking and utilize guidance strategies that command the missile along the line- of-sight (LOS) to the heavily armored areas of the tank. The target kill probability, P_k, for these systems is limited by the LOS guidance strategy and operator error. Therefore, new approaches including passive homing and non-LOS guidance are being developed to increase P_k for future antitank missile systems. Passive homing removes operator error during flight, and non-LOS guidance (ie: top attack) yields warhead impacts on the top areas of tanks which have less armor. A passive, top attack, antitank missile has many advantages; however, it relies greatly on a passive seeker to provide correct target state information under a variety of battlefield conditions. This paper introduces effects of missile-to-target range and line-of-sight (LOS) on passive seeker performance in battlefield conditions for a surface-to-surface missile (SSM) using an imaging infrared (IIR) seeker. Using this analysis, an adaptive missile guidance concept based on several guidance strategies is proposed to maximize P_k. Fundamental considerations and challenges for the proposed missile guidance concept are mentioned.

The Naval Research Laboratory (NRL) has designed and fabricated a prototype seeker system with an Infrared Focal Plane Array (IRFPA) sensor and custom image processor. This prototype seeker system, developed and used as flyable IRFPA anti-ship cruise missile threat simulator to test ship based IR countermeasure systems has improved performance over existing anti-ship systems using scanning IR sensors and image processing hardware. This paper addresses the enhanced performance of the IRFPA sensor and the custom image processor system, both separately and in combination. A comparison of the specifications and performance of a modern IRFPA versus existing scanning sensor is performed to illustrate the benefits of staring arrays in IR guided systems. In addition to utilizing a compact IRFPA sensor, the prototype seeker system takes advantage of improvements in image processing hardware, general purpose CPUs, multiprocessing, multitasking operating systems. An automated nonuniformity correction (NUC) calibration algorithm for the IRFPA sensor will be described which allows in situ calibrations without external calibration sources or human operator intervention. The image processing hardware consists of three 68020 CPU cards running in parallel under the real-time operating system pSOS+m, a convolution card, a histogram card, and an image analysis card. The 68020 CPUs handle tasks controlling the image processing, target classification, tracking algorithms, and data collection. The seeker software accumulates a history of frame parameters, classifies potential targets, and commands the gimbal electronics to point the sensor. The tracking algorithm can maintain history and track of multiple targets. Field test results show that the added sensitivity and resolution of the IRFPA and increased processing capability of the seeker system provide increased maximum track range, more robust tracking, and countermeasure hardness.

At Sandia Labs' Coyote Canyon Test Complex, it became necessary to develop a precision single station solution to provide time space position information (tspi) when tracking airborne test vehicles. Sandia's first laser tracker came on line in 1968, replacing the fixed camera technique for producing trajectory data. This system shortened data reduction time from weeks to minutes. Laser Tracker II began operations in 1982, replacing the original tracker. It incorporated improved optics and electronics, with the addition of a microprocessor- based real-time control (rtc) system within the main servo loop. The rtc added trajectory prediction with the loss of adequate tracking signal and automatic control of laser beam divergence according to target range. Laser Tracker III, an even more advanced version of the systems, came on line in 1990. Unlike LTII, which is mounted in a trailer and must be moved by a tractor, LTIII is mounted on its own four-wheel drive carrier. This allows the system to be used at even the most remote locations. It also incorporated improved optics and electronics with the addition of absolute ranging, acquisition on the fly, and automatic transition from manual joystick tracking to laser tracking for aircraft tests.

The operations or independent small, lightweight infrared (IR) sensors, called Notifiers and Interrogators, are coordinated to detect and track theater ballistic missiles (TBMs) during boost phase from high altitude aircraft. The Notifier provides early detection of TBM launches over a wide field-of-regard facing the earth. When a target candidate is detected, an Interrogator immediately points at the object to first validate the persistence of the object, to perform TBM authentication, and to determine trajectory coordinates of the TBM. The Notifier and Interrogator sensors differ in their optical trains but otherwise are identical; the later offering nearly 3.5x higher angular resolution and a filter wheel. The sensors must rapidly assess their fields-of-regard to accomplish their roles. A unique three-axis stabilized optical tracking platform has resulted from the requirements thus imposed on angular rates, settling times, and angular stability. The line-of-sight of each earth facing sensor is pointed by a step-and-settle mirror drive that covers a hemisphere field-of-regard including nadir and the horizon. The line-of-sight is steerable through 360 degrees in azimuth and 90 degrees in elevation. This paper describes the 'proof of concept' implementation and it provides details about he designs of critical components that make up the optical tracking platform and the step-and- settle mirror drives.

As the system engineering process flows down constellation coverage specifications to the Spacecraft level in terms of agility requirements it's critical that the relationships between manueverability and cost are clearly understood. The probability of optimizing the cost of typical ATP system would be greatly enhanced if a realistic integrated cost/engineering model were available during the initial phase of a program (e.g. Conceptual Design Phase). Most Cost Engineering work performed to date has been done by Cost and/or Systems Engineers which has typically lead to models with a cost emphasis. This work tends to be parametric in nature and hence the models have has little 'buy-in' from the design engineering side of the house. A better approach is to take existing credible engineering models for the key Spacecraft subsystems (Attitude Control, Thermal, Power, etc.) and to append these models to include the appropriate hardware databases. This would allow the models to output cost, power and weight, besides analytical engineering parameters like torque, momentum, etc.. For sound engineering reasons some, but not all, subsystem models should be time-domain based (dynamic) simulations--a clear diverges from the typical Systems Engineering approach. A modular spacecraft model like the one created at Lockheed for the FEWS/ALARM programs provides an ideal basis for developing a Dynamic Integrated Cost & Engineering (DICE) Model. This paper provides a 'snapshot' of the initial development of Attitude Determination and Control portion of the DICE Model. These subsystems were modeled first since maneuverability has such a large cost impact on them. A multiple body dynamics package, High TEC¹, provides the core of this DICE module. This package has been integrated into several simulation packages as described in previous works. Having access to this detailed 3-axis simulation model allows one to properly size spacecraft attitude systems (especially sensors and actuators). Hardware models can be easily interchanged through the friendly Graphical User Interface integrated with High TEC. The DICE Control system module outputs hardware attributes such as type (Reaction Wheels, Control Momentun Gyros, etc.), size, power, cost, etc., as a function of maneuverability. With this tool, Design and Systems Engineering can work together to optimize ATP cost rather than having coverage requirements flown down with little regards to design constraints and capabilities.

The present paper brings consistent physical (technical) rationales and mathematical method enable one to perform concise analysis of an effectiveness of hypothetical multiple element Randomized Anti-Missile Defense (RAMD) describing the offensive asset(s) (target(s)) penetrativity in terms of coupled probabilistic functionals taken along arbitrary 3D trajectories of all targets (attack profile) and conditioned on a set of dominant technical parameters and spatio- temporal dynamical variables. The key feature here is that analysis is carried out not only in terms of varying numbers of targets and RAMD- elements, but it also takes into account the evolving spatio-temporal (i.e, 4D) distributions of both adversaries. Paper provides an exposition of what temporally balanced quasi-optimal target-assigning strategy is and how the upper limit to target penetrativity depends on both generalized technical parameters and integral overlapping of continual element-density distributions and properly 'broadened' pathway of a target. The developed model makes it possible to analyze both purely defensive situations and missile (nuclear) exchange scenarios. It is also applicable to RAMD effectiveness issues exploration in cases of a massive raid detterence (Advanced Global RAMD) and protection against limited strikes (Point/Area RAMD), including direct attack on RAMD itself, whatever their sources may be. The study give general recommendations on transfer from analysis to a synthesis of imitative 'game' model version touching on battle management fragments. Formalized model offers scope for extension coming from further specification of nodal technical and environmental characteristics. The outlined abstract math model of effectiveness falls within the hypothetic RAMD. In attempt to treat this intractable problem the author relied on general scientific principles. The views expressed are those of the author.

Performance evaluation is fairly straight forward when tracking a single bright target. Tracking a dim target in clutter, on the other hand, can present some challenges in performance evaluation. For example, a target track can be a false track. A false track is one based primarily on clutter points or false signals rather than the target. How is a false track to be identified and what measures of performance should be used to account for a false track? In comparing two different trackers or two variants of the same tracker, how can the performance be evaluated effectively and fairly? One tracker might acquire the target earlier (at greater range) than the other, yet not track as accurately. In some scenarios one tracker might even miss the target completely, that is, exhibit a missed track. This is another performance evaluation issue that must be addressed. This paper presents a two-step method for evaluating performance of tracking a single target in clutter. The first step classifies the target track as valid, missed or false. Various tracking measures of performance are then computed in the second step. This two-step approach provides a systematic method for tracker performance evaluation. More importantly, this methodology is designed to permit a fair evaluation and comparison oftwo or more competing trackers. KEYWORDS: Performance Evaluation, Measures Of Performance, Trackers, Tracking, Single Target Tracking, Multiple Object Tracking, Multiple Track Processing, Multiple Sensor Tracking, False Tracks.

This paper investigates the problem of tracking highly dynamic targets in real-time. The advantages and disadvantages of various control and interpolation techniques are discussed. Algorithms appropriate for real- time tracking are compared with algorithms that are common in applications with a priori knowledge of the trajectory. A unique command interpolation technique combined with a state feedback control structure, implemented on a commercial DSP controller, demonstrates a significant performance improvement.

The Transportable Algorithm TestBed (TATB) was designed as a software framework to evaluate different algorithms and processing modules used for target detection, tracking, and signature extraction. The processing chain has been divided into 'pixel processing' involving pixel-by-pixel corrections, such as camera gain and offset corrections, 'object processing' which is single frame object detection and characterization, radiometric and angular calibration, and tracking or 'frame-to-frame correlation', in which multiple objects are associated across frames in an image sequence. The TATB framework is such that one can select a particular combination of processing modules to link together, forming a complete processing chain, and evaluate the detection, tracking, and signature extraction capabilities of that particular combination of processing modules. This paper will discuss the operations of one particularly robust set of processing modules and present results obtained using those modules. The TATB system was also designed to be able to connect to different data sources, either real-time digital image sequences, or data stored in disk files in various formats. The ability to accept different format data has been tested and found to be extremely flexible. The TATB system has been applied to defense, wildlife tracking, and commercial applications.

An extended rule set multitarget tracking system based on fuzzy logic is discussed. The system that we describe utilizes fuzzy logic principles in a Kalman-type fuzzy tracking filter for the purpose of drastically reducing computation overhead. The tracking problem is modelled as a discrete time system in the standard state form. Rather that employing the Kalman filter, an extended rule set is used at each iteration to update an estimate of the target state vector. The update is based upon the innovation vector for the target, which is the vector difference between observed and estimated target position. Since each target has many returns, both actual and clutter, it is necessary to validate each return before including it in the calculation of the average innovation vector. Return validation is accomplished via fuzzy logic and is incorporated with a fuzzy measure of target-return similarity in the fuzzy return processor (FRP) to obtain the average innovation vector. Spurious returns are discussed in terms of their causes and their effect on the fuzzy tracking algorithm. The rule set is described in detail, and the results are discussed and compared to Kalman filter tracking results. Finally, the algorithm is applied to an actual forward-looking infrared (FLIR) image sequence provided by Texas Instruments, Inc. and results are discussed accordingly.

In this paper, two new centralized tracking systems are proposed. The first system employs a fusion center to preprocess multisensor measurements. In the fusion center, a novel multisensor track initiation processor based on the logic-based initiation concept is used to determine new tracks. The output of the fusion center are fused measurements which are sent to a multitarget tracker, such as nearest neighbor or joint probabilistic data association filter, to produce the global state estimates. The second centralized system is a fully centralized tracking architecture in the sense it does not use fusion as a preprocessor as in the first case. In this system, multisensor data association and multisensor track initiation are used with a standard Kalman filter to perform multisensor tracking. Both centralized tracking systems are evaluated in various tracking environments. Comparisons between the two systems indicate that the first system is more efficient in eliminating clutters. The second one requires fewer scans to initiate tracks and has better performance and a lower computational complexity.

A VME based real-time control system has been developed for use in the testing of smart munition weapons systems. The testing of advanced multimunition systems requires a platform that has not only a robust and stable servo control loop, but also a data collection platform that is capable of acquiring and tagging a wide range of sensory data. The data collection scheme must be able to handle synchronous, asynchronous, and multiframe rate sensor inputs and be capable of handling changing modes of operation in real-time. To meet these requirements, a DSP platform was utilized for the servo control loops, while programmable hardware logic was utilized to allow deterministic strobing of the time and pointing information. Discussions of the imaging requirements for this application, and limitations and uncertainties involved with optical tracking measurements are presented.

The Naval Research Laboratory in Washington, D.C. has been conducting extensive research in the area of infrared tracking systems. Automatic detection of sea surface targets has been experimentally investigated by using a programmable passive infrared seeker system. The Victor seeker system uses two independent programmable infrared tracking systems which are mounted on naval P-3 aircraft and flown at sea surface targets. This system contains an infrared scanning detector array assembly, an infrared camera imaging system and a VME multicomputer control/image processor. The seeker system permits varying key parameters in software and evaluating the performance in real world target scenarios.

The Navy has a need to score Naval Gunfire Support exercises using a virtual land mass, positioned off the Pacific Missile Range Facility (PMRF) at Barking Sands, Kauai, Hawaii, as a target. The target area is approximately 15,000 yards from shore. As a part of this exercise, illumination rounds that detonate 1000 to 1500 feet above the surface and airburst high explosive rounds that detonate approximately 50 feet above the surface must be scored. The design requirements for the measuring system for the illumination rounds is detonation position accuracy in three dimensions of approximately +/- 15 yards and for the high explosive airburst rounds the determination of air as opposed to water detonation. Detection and measurement is to be automatic. The optical portion of the system consists of two shore based high resolution CCD cameras mounted on precision rotators, surveyed in via GPS and supported by appropriate hardware and software. One station is located 80 feet above sea level and the other is located 1500 feet above sea level. Data from both stations is telemetered to a central station for position computation and display. The base line along the shore is approximately 10,000 yards. Preliminary measurements, system design, error analysis, and calibration techniques will be discussed.

The Aspect Camera Assembly (ACA) is a state-of-the-art star tracker that provides real-time attitude information to the Advanced X-Ray Astrophysics Facility-Imaging (AXAF-I), and provides imaging data for post-facto ground processing. The ACA consists of a telescope with a charge coupled device (CCD) focal plane, associated focal plane read-out electronics, and an on-board processor that processes the focal plane data to produce star image location reports. On-board star image locations are resolved to 0.8 arc-sec, 1 (sigma) and post-facto algorithms yield 0.2 arc-sec, 1 (sigma) star location accuracy. A high- fidelity ACA breadboard was built along with a high accuracy test facility. In air, image position determination has been verified to 0.2 arc-sec, 1 (sigma) accuracy. This paper presents the system requirements and a description for the ACA. A design description is included for both the ACA and the test facility, and performance results are presented.

This paper deals with the restoration of images blurred as a result of image motion or vibration. The key for restoration algorithm success is to derive accurately the Optical Transfer Function (OTF) representing the image motion degradation in the spatial frequency domain. The basic method of obtaining the OTF from the relative displacement between the camera and the object using a motion sensor has been developed recently and is discussed elsewhere. In this paper, the motion function is derived instead from analysis of a sequence of images. The first step is to obtain the image motion information from the sequence of images according to two well known algorithms - the Block Matching Algorithm (BMA) and Edge Trace Tracking (ETT). The basis for these two methods consists of tracking a block or an edge through a sequence of several images. The results of these two methods were fitted to a sinusoidal function, compared, and there was excellent agreement between them. Finally, the image is restored using the OTF obtained from the tracking method.

Multiple target tracking algorithms are computationally intensive even for a small number of targets because of complexity of kalman filter and Data Association techniques. Joint Probabilistic Data Association technique is a robust MTT algorithm for tracking dense targets in cluttered environments, with N update the MTT algorithm using ellipsoidal gate and JPDA for data association for maneuvering targets has been implemented on DSP chip TMS 320C25 exploiting its architecture and instruction set.

A high priority Air Force interest is to understand the tilt (or jitter) induced by atmospheric turbulence on a laser beam propagated over a high altitude long horizontal path. This will aide in the design of the pointing and tracking systems such as used for communication systems and laser beam propagation. The Phillips Laboratory has undertaken a technology program to simulate, measure, and evaluate the implications of this high altitude tilt. This paper includes 3 related efforts; a simulation that predicts high altitude tilt using present knowledge of the atmosphere, a flight measurements program that is gathering data at high altitude, and a simulation effort to understand the implications on a pointing and tracking system of this high altitude tilt. The simulation is a ray trace code that has been developed at the Phillips Laboratory over the past 15 years and has been anchored using available data. Until recently this simulation had not include high altitude horizontal propagation data. However, the Phillips Laboratory has recently completed a high altitude measurement of scintillation and a temperature probe measurement effort to measure the atmospheric structure. This information has been included in the simulations. A controls simulation is used to evaluate the requirements for bandwidths and controls design and utilizes the disturbance data produced by the atmospheric simulation.

The High Altitude Balloon Experiments (HABE) control architecture design focuses on establishing an inertial stabilized line-of-sight (LOS) for the tracking and laser pointing subsystems. High bandwidth LOS stabilization is implemented with an inertial reference measurement system. The Inertial Pseudo Star Reference Unit (IPSRU), and inertially stabilized two degree of freedom platform, generates an inertially stabilized alignment reference beam which probes the multiple aperture system. Fast steering mirrors (FSM) in optical alignment loops track the alignment reference beam performing jitter stabilization and boresight alignment. The auto alignment system operates in the primary aperture beam path, stabilizing the fine tracking sensor imagery and surrogate high energy laser pointing subsystem. Due to the superior performance of the IPSRU stabilization platform, aggregate LOS stabilization system base motion and optical jitter rejection is directly traceable to the auto alignment system control dynamics and sensor noise performance. Performance requirements specify two axis FSM control bandwidths of 500 Hz with a positioning resolution better that 300 nano-radians in output space. The digital control law is implemented in high performance digital processors with sample rates in excess of 15 kHz. This paper presents the bench top integration and testing of the digital auto alignment system beginning with a discussion as to the reason behind choosing a digital implementation, a opposed to a much simple analog implementation. A description of the error budget requirements of the HABE digital auto alignment loop follows. The components comprising the auto alignment loop, including mirror and processor hardware and software are described. Experimental objectives are presented with a description of the laboratory setup. Simulation models are constructed from component test data to aid in the development of the alignment system control architecture and discrete time control law realizations. The experimental data and methods for testing of the real time implementation used to optimize the controller design and evaluate auto alignment system performance are discussed. The paper concludes with lessons learned and a discussion of future HABE program work concerning LoS stabilization and the implementation of high bandwidth digital control system architecture.

A gyroscope star calibration algorithm developed for use in the High- Altitude Balloon Experiment, a program to resolve acquisition, tracking, and pointing/fire control issues in support of future directed energy programs, is presented. The initial acquisition of a target from a free- floating balloon platform necessitates an inertial pointing capability. Gyroscopes are normally used to measure angular rotations with respect to an inertial reference frame. Given an initial attitude, the gyro measurements can be used to determine the attitude at any later time. For the High Altitude Balloon Experiment, GPS and a magnetometer are used to provide an initial estimate of the payload attitude. This paper describes a deterministic algorithm, which uses star sensors measurements to correct initial attitude uncertainty errors. In addition, a simple method for measuring gyro bias in flight is presented. The consequences of star sensor to gyro frame misalignment are examined for both rate gyros and angular pulse output gyros. This paper describes the algorithm used to meet the calibration accuracy required to inertial pointing of the gimbaled payload.

The rapid development of digital processing raises the possibility for high speed control system architectures which are distributed in nature; that is, high speed special purpose processors tied together with a centralized overall executive. This processing architecture, while allowing more flexibility to the software programmer, introduces additional complexity into the control system design. For example, the separate digital blocks will be subject to 'slightly random' timing jitter which may degrade control loop performance and stability. This paper applies H_(infinity ) analysis techniques to a multirate/asynchronous sampling digital control example to determine stability and control performance.

The purpose of this paper is to discuss the daytime tracking of astronomical objects. We are interested in this because many stars are not visible in the night sky, or we may want to perform around the clock observation of other astronomical objects. Daytime tracking presents many difficulties, including high sky background levels and low object to background contrast. In this paper I will describe a daytime tracker experiment, give a description of the various components, and talk about the bandwidth requirements and performance of the system. I will discuss the contrast and background problems as well as the use of video processor boards to mitigate these problems.

This paper describes a new observer design method that allows for estimating the angular rates along a vehicle's three principal axes. This method uses measurements from a single two-axis angular rate sensor (gyro) and determines the rates for the third axis by using a nonlinear observer. Unlike conventional approaches where the equations governing vehicle motion (Euler's equations) are linearized and then an observer is constructed based on the linear model, this method does not require linearization of the system. Instead, a pseudo-linear representation is used. The pseudo-linear model is obtained by systematically decomposing a nonlinear system into linear and nonlinear terms. The nonlinear components are then redefined as an auxiliary set of state variables and/or inputs. This leads to an augmented linear system representation that is mathematically equivalent to the original nonlinear system. This method allows standard linear observer design techniques to be applied, and it develops observers that are capable of estimating the third-axis angular rates using measurements corresponding to the other two axes. The method's effectiveness is illustrated with an example. The case studied is the complete attitude rate determination and control of a spinning spacecraft. Computer simulation results show that the new approach provides excellent three-axis attitude control, yet requires angular rate sensors for only two axes.

This paper describes the design, development and performance of a lightweight precision gimbal with dual-axis slew capability to be used in a closed-loop optical tracking system at Lawrence Livermore National Laboratory-LLNL. The motivation for the development of this gimbal originates from the need to acquire and accurately localize warm objects (T approximately equals 500 K) in a cluttered background. The design of the gimbal is centered around meeting the following performance requirements: Pointing Accuracy with control < 35 (mu) rad-(1-(sigma) ), Slew Capability > 0.2 rad/sec, Mechanical Weight < 5 kg. These performance requirements are derived by attempting to track a single target from multiple satellites in low Earth orbit using a mid-wave infrared camera. Key components in the gimbal hardware that are essential to meeting the performance objectives include a nickel plated beryllium mirro, an accurate lightweight capacitive pickoff device for angular measurement about the elevation axis, a 16-bit coarse/fine resolver for angular measurement about the azimuth axis, a toroidally wound motor with low hysteresis for providing torque about the azimuth axis, and the selection of beryllium parts to insure high stiffness to weight ratios and more efficient thermal conductivity. Each of these elements are discussed in detail to illustrate the design trades performed to meet the tracking and slewing requirements demanded. Preliminary experimental results are also given for various commanded tracking maneuvers.

High accuracy pointing performances and also large autonomy from the host satellite are required for the future optical payloads designed to perform astronomy, laser communication, planetary observation, or other missions. In order to prepare these future applications, SFIM Industries/Ets d'Asnieres (SFIM/EA) has undertaken under CNES (French Space Agency) contract, the development of a new generic system able to fit a large variety of satellites, instruments and missions. To meet the stringent pointing stability constraints, a few microradians, the design has mainly to minimize friction and to control flexible modes. The mechanism is a two axis single stage turret designed around two identical joints, linked together by an isostatic truss structure. The structure is optimized using an integrated design approach according to the flexible modes control problem. The high angular stiffness to friction ratio of the articulation is obtained through an optimized design of the ball bearings. These are protected during launch by a new locking mechanism. The residual friction wires torques are linearized by a mechanical device which provides in addition a closed loop friction compensation. The control torques are produced by a brushless torque motor optimized for low ripple and detent torques. The control laws design is based upon H(infinity) methods which have been adapted to robust control of flexible systems. The control loop uses a combination of optical encoder and accelerometer outputs or in a best way a low frequency payload ecartometry signal combined with accelerometers outputs. A feasibility study was carried out in 1993 and has demonstrated the validity of the concept. A demonstrator is to be tested in the end of 1995.

The requirements concerning optical instruments stability on satellites are getting more and more stringent. At the same time, those instruments have to put up with a disturbed dynamic environment caused by other on- board equipments which can be made worse by flexible structures. A possible design option for achieving pointing stability can consist in isolating the optical sensitive payload from the host satellite. Such a concept is being studied by CNES (french space agency) on a small angular range (+/- 5 degree(s)) payload and a laboratory breadboard of a two axis gimbal 'softmount' system has been developed using flex pivots to carry out passive isolation. The instrument's Line-of-Sight (LOS) stabilization is performed through a control algorithm using acceleration feedback and positive feedback on relative position error in order to decrease the flex pivot resonance. These informations are determined respectively by angular accelerometers and capacitive encoders mounted on each axis. Another design option consists in a 'rigid mount'. In this case the LOS stabilization is performed by a strong control on inertial position error information. This alternative is being studied using a specific optical bench to provide the inertial position error. These two concepts are limited in performance and an optimal solution has to combine both principles to achieve the demanding pointing stabilization requirements of future satellite systems. This paper presents primarily first experimental results and highlights problems encountered, such as delay or flexible modes, as well as solutions proposed.

On the basis of analyzing and proving the circular symmetry of circular beam deflectors this paper presents a new line-of-sight stabilized prism which consists of four roof-Dove prisms. The characteristics of the new prism is that its volume is small and its structure is rational and that the circular symmetry of the new prism is more evidently shown.

Track maintenance refers to the process of fusing sensor contacts to existing tracks in order to estimate the target state. Target tracking in cluttered environments is difficult because there can be several contact-to-track associations for a given track. The Nearest Neighbor (NN) approach, which uses the contact nearest the predicted measurement, is an early maintenance algorithm for clutter but its performance greatly suffers because alternative hypotheses are not considered. Probabilistic methods, such as the Probabilistic Data Association (PDA) and Integrated PDA (IPDA), are superior because all hypotheses are considered. The PDA inherently assumes that the track exists where as the IPDA does not. A single PDA or IPDA filter has difficulty with maneuvering targets and will therefore be incorporated into the Interacting Multiple Model (IMM) algorithm. For illustrative purposes, simulation studies will be used to compare the performance of the algorithms.

Since phased array radars have the ability to perform adaptive sampling, proper radar control can improve many aspects associated with the tracking of multiple maneuvering targets. Techniques have been developed to significantly reduce the sampling of a maneuvering target in which the savings in the radar time-energy budget can be reallocated to accomplish other functions. However, controlling the revisit time becomes more difficult when the target is maneuvering in the presence of false alarms. The technique proposed in this paper used the Interacting Multiple Model (IMM) algorithm and Integrated Probabilistic Data Association Filter (IPDAF), combined to form the IMM-IPDAF, to track maneuvering targets and control the radar revisit time in the presence of false alarms.

This paper discusses the problem of registration which is a prerequisite process of a data fusion system to accurately estimate and correct systematic errors. An exact maximum likelihood (EML) registration algorithm is presented. The likelihood criterion is formulated by transforming the measurement data from local sensors to a common system plane. The algorithm is implemented by applying a recursive two-step optimization which involves a modified Gauss-Newton procedure to ensure fast convergence. Numerical simulation studies are conducted to show the effectiveness of the algorithm and comparisons with other registration approaches are provided.

Probabilistic data association (PDA) is an important class of data association filters. This paper analyzes the performance of four types of PDA techniques: the standard PDA, the joint probabilistic data association (JPDA), the cheap JPDA and sub-optimal JPDA. Both the weighted and nearest neighbor PDA filters are evaluated and their track quality compared using real radar data sets which contain air targets of different characteristics, collected in various conditions.