NASA Goddard Space Flight Center and AppNet, Inc. are developing a general and highly extensible framework that applies to any kind of instrument that can be controlled by a computer. The software architecture combines the platform independent processing capabilities of Java with the power of the Extensible Markup Language, a human readable and machine understandable way to describe structured data. A key aspect of the object-oriented architecture is software that is driven by an instrument description. This description is written using the Astronomical Instrument Markup Language, a domain specific implementation of the more generalized Instrument Markup Language (IML). IML is used to describe graphical user interfaces to control and monitor the instrument, command sets and command formats, data streams, communication mechanisms, and data processing algorithms.

The SOLIS (Synoptic Optical Long-term Investigations of the Sun) project is constructing new telescopes to replace the Vacuum Telescope at Kitt Peak. Among other goals, SOLIS is to continue observations that have been in progress over the last 25 years, using a new set of instruments. Solar observing differs from other astronomical observing in several key ways: exposures tend to be short and yet still produce large volumes of data--the expected data rate is a sustained 56 MBytes per second with peaks of 132 MB/s. The overwhelming majority of observations are synoptic although the ability to respond to unscheduled events (e.g. solar flares) is a cornerstone of the system. Precise time constraints are typical observation requirements.

The Very Large Telescope (VLT) Telescope Control Software (TCS) is a portable system. It is now in use or will be used in a whole family of ESO telescopes VLT Unit Telescopes, VLTI Auxiliary Telescopes, NTT, La Silla 3.6, VLT Survey Telescope and Astronomical Site Monitors in Paranal and La Silla). Although it has been developed making extensive usage of Object Oriented (OO) methodologies, the overall development process chosen at the beginning of the project used traditional methods. In order to warranty a longer lifetime to the system (improving documentation and maintainability) and to prepare for future projects, we have introduced a full OO process. We have taken as a basis the United Software Development Process with the Unified Modeling Language (UML) and we have adapted the process to our specific needs. This paper describes how the process has been applied to the VLTI Auxiliary Telescopes Control Software (ATCS). The ATCS is based on the portable VLT TCS, but some subsystems are new or have specific characteristics. The complete process has been applied to the new subsystems, while reused code has been integrated in the UML models. We have used the ATCS on one side to tune the process and train the team members and on the other side to provide a UML and WWW based documentation for the portable VLT TCS.

Last two years the GTC control system has continued its design and prototyping activity. This paper presents the current state of the GTC control system architecture. The overall architecture is already designed and the main subsystems have been specified and are being designed. Design patterns have been applied to the architecture design and are also used as an effective way to document the design.

A Rapid Prototype and full development plan of the SOAR TCS is reviewed to show advances in: (1) Prototyping speed, which makes implementation and test of features faster than specification under older methods. This allows the development environment and prototype modules to become partners with and part of the specification documents. (2) Real-Time performance and reliability through use of RT Linux. (3) Visually Rich GUI development that allows an emphasis on `seeing' versus `reading'. (4) Long-Term DataLogging and Internet subscription service of all desired variables with instant recall of historical trend data. (5) A `plug-in' software architecture which enables rapid reconfiguration and reuse of the system and/or plug-ins utilizing LabVIEW graphical modules, a scripting language engine (in LabVIEW) and encapsulation of interfaces in `instrument-driver' style `plug-in' modules. (6) A platform- independent development environment and distributed architecture allowing secure internet observation and control via every major OS and hardware platform.

A Ritchey-Chretien 1.2 m telescope (EULER) and the High- Resolution echelle Spectrograph (CORALIE), a new Swiss observing facility at ESO La Silla Observatory, are operational and since Summer 1998. The Observatory operation is fully automated and supports the unattended, attended and interactive mode of operation under local or remote control. The control hardware is based on Local Control Units (LCU) built from PC/RedHat Linux computers and a Unix Computing Server. The Operational Software is built around INTER, a command language interpreter featuring communication control, data access, image processing functions and easy access to external resources. The general SW architecture is a non-hierarchical tree of pairs made of hardware- independent interpreters running on the Observing Server and hardware dependent servers running on the LCU's. The Operational Software includes the full access (Creation/Modification/Retrieval) to the input/output databases, telescope, instrument and auxiliary set-up and control files as well as a full data-reduction pipeline. We describe briefly the system architecture, summarize the performances and the experience gained over 18 months of operation and we discuss some critical issues: use of standard components, parallel operation, real-time requirements, system upgrade and maintenance.

The WIYN Tip-Tilt Module is being developed to improve delivered image quality over a 4 X 4 arc-minute field across the B through H bands. This paper details the software implementation of the project under real time Linux and LabVIEW.

The Telescope Nazionale Galileo (TNG) telescope is now operational. One of its main goals is to provide high quality images, in a wide range of operating conditions and for several observing modes. Telescope pointing and tracking performances can heavily affect achievement of this requirement, and particular care must be taken in order to reach the highest possible accuracy. Control of the three axes is implemented in one VME controller (minimizing data exchange through the TNG LAN), and telescope mount positions are computed from object coordinates taking into account physical and environmental aspects which alter the object apparent position. Improvement of telescope pointing and tracking performances is obtained by two means. Systematic errors are mostly corrected using a model compensation, that can be introduced in the coordinates transformation flow. The telescope model is derived by off-line analysis of the pointing errors on a specific set of data. In this way we could improve the pointing and tracking performances up to now by a factor of about 30. Tracking drift given by non- systematic and residuals of systematic errors is corrected using a guide camera, which mounts a 800 X 576 CCD (0.35 arcsec/pixel scale). Guide stars are selected from an on- line available star catalogue. Light from the selected star is focused on the guide camera by moving a probe housed in the rotator-adapter module. Tracking drift computation is performed on a two-axes scheme (Right Ascension and Declination), with sub-pixel accuracy. Here the first results of the telescope pointing and tracing accuracy are presented.

We present the current status of the antenna control software for the Submillimeter array. This software is responsible for pointing and tracking astronomical sources and four antenna pointing calibration. We describe the various stages of the calculations, starting with source- lookup from catalogs and resulting in antenna coordinates commanded to the servo computers. We also present some preliminary results on the pointing calibration of the antenna mounts using an optical guide-scope.

This paper presents the Green Bank Telescope (GBT) antenna control system architecture, describes the tracking options available to the observer, and details the GBT pointing, focus tracking and active surface control systems, as well as approaches to vibration reduction by use of minimal jerk trajectory generation. A description of the phased approach to improve telescope performance through the use of laser metrology range finders is given.

The Field Stabilization functionality compensates for image motions, mainly due to wind gusts, with a frequency of up to 50 Hz. Position errors are detected with a CCD mounted in the guide probe, and the correction vectors are passed to the secondary mirror, which can perform tip/tilt corrections with a frequency up to 100 Hz. This functionality is in regular use on the first two VLT telescopes, Antu and Kueyen; on Antu in operational use since 1st of April 1999 and on Kueyen for commissioning tests. In fact, Field Stabilization is ALWAYS used during observations. It is started as an integral part of the setting of the telescope. This paper describes the software and hardware that performs the Field Stabilization functionality. The architecture of the system, both software and hardware, is presented, with a discussion about problems and special solutions. The actual status, including some results from the commissioning and operational phases of the two first telescopes, is described. Finally, we present a short discussion about possible future extensions and improvements.

The advent of new large CCD array cameras necessitates computing systems that are highly optimized but also enable flexibility of operation and ease of programming. We describe `UltraDAS', a CCD-control and data-acquisition system for the Isaac Newton Group of Telescopes that achieves these aims by combining high-performance detector- controllers with modern programming techniques on current UNIX workstations.

The Cassegrain Instrument Automatic eXchanger (CIAX) system for the 8.2 meter Subaru Telescope moves instruments between the Cassegrain mounting flange and stand-by flanges without manual intervention. Observation efficiency improves not only because of quick exchanges, scheduled or emergency, but also because of increased flexibility in selecting an optimum instrument for weather conditions or observation goals. Reliable and safer instrument exchanges are achieved by the precision mechanical positioning system (less than 0.5 mm) and an automatic connector system for electrical cables, optical fibers and fluid lines. Instrument down time due to connector/cable failure by human error is eliminated. Interfaces to the telescope flange are standardized for all five Cassegrain instruments (approximately 2000 kgf each) currently in use or under preparation.

This paper is focused on how the features and object services provided by CORBA (Common Object Request Broker Architecture) can be applied to solve the specific needs of a modern telescope control system. The solutions presented in this paper are based on TAO, which is the ORB (Object Request Broker) implementation selected by the GTC Control Group, although any ORB implementing OMG's CORBA 2.3 standard could be used.

The National Solar Observatory's SOLIS project is building a new solar telescope to replace the existing Vacuum Telescope at Kitt Peak, Arizona. Three new solar instruments will replace existing capabilities and provide new scientific capabilities over the expected 25 year lifetime. The SOLIS instruments are the VSM (Vector SpectroMagnetograph), FDP (Full Disk Patrol), and ISS (Integrated Sunlight Spectrometer). Each provides a unique solar physics science capability for magnetic, image, and spectral data. Part of the challenge for the SOLIS software has been controlling and coordinating these varied instruments to work efficiently in a synoptic observing environment.

The SOFIA data system must meet numerous technical and organizational objectives, including widely available distribution to support integration and testing at users' institutions. As with all professional data system software development, a wide range of sophisticated development tools are required. With open source software now widely available, it is possible to build an advanced Unix-based development environment taking full advantage of freely available tools. This paper analyzes advantages and disadvantages of this approach, the selection processes used, and the list of tools selected to date for the SOFIA development effort.

The application of computers and software to control continuously or on demand safety critical processes is unavoidable in complex telescopes and associated operation and maintenance support equipment. While the control of the maximum azimuth and elevation movement ranges of a telescope is a rather simple task and hardwired high reliable endstops can be used to practically eliminate any risk associated with the potential exceeding of the range limits, the handling and recoating of a 8 m thin Zerodur mirror is quite a delicate activity. It includes already a number of computer controlled actions which--if no risk reduction measures are established--lead to very high risk. But in complex systems like an 8 m optical telescope, coating facilities or similar the functional control and the control of major safety critical operations have to be performed by computers. Straightforward hardwired interlocks devices are not adequate for the control and monitor of highly complex processes. The still increasing process complexity will also lead to the demand for fast and intelligent risk limitation and reduction devices, and the use of safety related systems based on electrical, electronic and/or programmable electronic technology is unavoidable. But in using computers and software in safety critical applications the applicable standards and norms have to be applied strictly.

This paper gives details about our experience with the development and installation of the Very Large Telescope (VLT) control software, considering standardization, iterative development, release concept, testing, configuration control, which were all elements of our approach.

This paper presents the actual status of the Telescope Control Software on the VLTs and on the other telescopes. The main focus is in the characteristics that make the TCS architecture portable on very different mechanical and optical configurations. Then the paper concentrates on the strategy used throughout the projects of integration and testing of the modular components. Particular attention is given to the installation and commissioning phases for the VLT telescopes.

This paper presents the distributed hardware and software architecture of this 1.2-meter telescope control system, entirely designed and built by the Geneva Observatory. The modular concept and the choices of hardware tested in industrial automation made it possible to obtain great operational robustness and guarantee for long-term maintenance. The adopted solution is based on a transputer tree-network. The interactions between telescope and observer are transparent and completely integrated in the observation software of the attached instrument.

The steady improvement in telescope performance at UKIRT and the increase in data acquisition rates led to a strong desired for an integrated observing framework that would meet the needs of future instrumentation, as well as providing some support for existing instrumentation. Thus the Observatory Reduction and Acquisition Control (ORAC) project was created in 1997 with the goals of improving the scientific productivity in the telescope, reducing the overall ongoing support requirements, and eventually supporting the use of more flexibly scheduled observing. The project was also expected to achieve this within a tight resource allocation. In October 1999 the ORAC system was commissioned at the United Kingdom Infrared Telescope.

Faint Object Camera And Spectrograph (FOCAS) is completed and now waiting for a commissioning run on the Subaru Telescope atop Mauna Kea. We have developed a software system that includes the control of FOCAS instruments, Multiple Object Slits (MOS) design, and an analyzing package especially for evaluating performances of FOCAS. The control software system consists of several processes: a network interface process, user interface process, a central control engine process, a command dispatcher process, local control units, and a data acquisition system. These processes are mutually controlled by passing messages of commands and their status each other. The control system is also connected to Subaru Observation Software System to achieve high efficiency and reliability of observations. We have two off-line systems: a MOS design program, MDP, and an analyzing package. The MDP is a utility software to select spectroscopy targets in the field of view of FOCAS easily through its GUI and to design MOS plates efficiently. The designed MOS parameters are sent to a laser cutter to make a desirable MOS plate. A special package enables prompt performance check and evaluation of the FOCAS itself during a commissioning period. We describe the overall structure of FOCAS software with some GUI samples.

This presentation describes our experience developing astronomical instrument control software for the Gemini 8m telescopes using the Experimental Physics and Industrial Control System (EPICS). EPICS originated in the particle physics community and is now being used widely in the astronomy community. The differences between the requirements and techniques of these two communities has meant that the development of astronomical instrument control software with EPICS has been a challenge. We explore the different methods used for astronomical instrument control software and describe the method chosen for the Gemini Multi-Object spectrograph (GMOS) software. We have developed an `assembly control' record to contain the high- level intelligence for each assembly and a `device control' record to control each assembly's individual mechanisms. The solutions developed by GMOS can be reused by other astronomical instruments, provided they have similar kinds of mechanism. We describe improvements that can be made to the GMOS records to make them more adaptable and propose the creation of a pool of EPICS solutions for the benefit of future instrument software developers. We also describe the future development of astronomical device control software in EPICS and propose a new hierarchical model.

All Cassegrain spectrographs suffer from gravitationally- induced flexure to some degree. While such flexure can be minimized via careful attention to mechanical design and fabrication, further performance improvements can be achieved if the spectrograph has been designed to minimize hysteresis and has active compensation for any residual flexure. The Echellette Spectrograph and Imager (ESI), built at UCO/Lick Observatory for use at Cassegrain focus on Keck II, compensates for such residual flexure via its collimator mirror. The collimator is driven by three actuators that provide control of collimator focus, tip, and tilt. The ESI control software utilizes a mathematical model of gravitationally-induced flexure to periodically compute and apply flexure corrections by commanding the corresponding tip and tilt motions to the collimator. In addition, the ESI control software provides an optional, manual, closed-loop method for adjusting the collimator position to compensate for any non-repeatable errors. Such errors may result from mechanical hysteresis or from thermally-induced structural deformations of the instrument and are thus not accounted for by the gravitational flexure model. This method relies on measuring the centroid position of fiducial spots within each echellete image. The collimator is adjusted so that the positions of these spots match those in a reference image. These spots are produced by a small round hole in the slit mask located near one end of the slit. We discuss the design and calibration of this flexure compensation system and report on its performance ont he telescope.

The Gemini Primary Mirror Control System is a distributed control system responsible for definition, figure and temperature control of the 8.1 meter diameter meniscus primary mirror of the Gemini telescopes. This paper describes the major design and implementation issues of the system and the experience to date of commissioning it.

The Controller Area Network (CAN), initially developed for the automotive industry, is becoming increasingly popular in industrial process control applications. The need for distributed low data rate monitor and control networking in industry is similar to the needs of the various instrumentation and support equipment in a modern radio telescope. In particular, immunity to noise and low radio frequency emission characteristics are common to both domains. The Atacama Large Millimeter Array project has adopted CAN technology for use in local monitor and control applications at each of its 64 antennas. A standard interface slave node providing flexible I/O options is under development and a simple application-level protocol making use of CAN to access these nodes in a master/slave fashion has been implemented. This paper will present the work which has been completed to date including experiences in the use of CAN in an astronomical environment. In addition, analysis and simulation of CAN networks is compared with the performance of our implementation in the lab.

POM, the `Pointing modeling module' of the VLT control system measures the pointing errors and derives a pointing model that is then applied within the tracking software of the telescope. The fit to the errors employs P.T. Wallace's TPOINT package, but POM is more that a wrapper for that software.

This paper describes `The Automotive Report Generator' used as engine to create daily and trend web reports about the telescopes and instruments operations at the Paranal Observatory.

Telescope, dome and camera controls can be seen as independent systems managed by an ad hoc software. We have used both hardware intelligence and a distributed PC based software to produce a system performing interactive and automatic observations. An integrated and automated data reduction pipeline allows most almost real-time image processing and WEB searchable archiving.

We are investigating the desired features of a new control system for the IRAM 30 m telescope, in particular new and improved observing modes for the millimeter wavelength range. Flexible and fast on-the-fly observing for spectral lines as well as continuum will play a central role. On-the- fly can be combined with frequency or wobbler switching, and will support focal-plane array receivers. To combine the goals for a system that is both flexible and easy-to-use, we want to use a feature-rich scripting language, combined with standardized commands or templates for the most common observing modes. Several recent developments within the current control system include ideas and solutions for the new control system; most of these run on Linux subsystems. Information about our plans and ongoing development can be found on the WWW by following links from the IRAM, Granada, home page.

As major observatories are planning automatic and optimized scheduling of large astronomical facilities, reliable and accurate monitoring of observing conditions is a pre- requisite. For this purpose, the concept of Astronomical Site Monitor has been developed for the VLT as an integrated sub-system of the observatory.

Optimization of observation sequences is a function necessary to get high efficiency and reliability of observations. We are now implementing scheduling software into the observation software system for Subaru telescope. The scheduling engine, SPIKE, developed at STScI is used with some modification for the Subaru telescope. An observation target list prepared by observers is converted to SPIKE Lisp codes. SPIKE output is inversely converted to Subaru commands to be executed with the observation software system. Real-time scheduling is planned to re-schedule observations by judging the weather and satisfaction conditions with the help of observation history. The scheduling software can be also used as support tools for observers to indicate an object good for the next observation.

The 3.6 m TNG (Telescopio Nazionale Galileo) Italian alt-ax telescope is now regularly working in La Palma, Canary Islands. The drive and control system of the main axes is working since October, 1997. It has been used for some months to support the installation of optics and other mechanical subsystems. Since June, 1998 the control system is integrated in the overall TNG informatic environment. In 1998 TNG has seen its first light. This paper reports the last most significant results for axes control, compared with the TNG tracking error requirements.

Space navigation and high performance instruments of observation (having a tiny field of view) demand higher speed and accuracy in computing orbital elements for natural or artificial near-Earth objects. The most important perturbations that act on these celestial bodies are: non- centrality of the gravitational field, Sun and Moon attraction, air drag and solar radiation pressure. Taking into account these perturbations and how they affect the orbital elements, we developed software based on semi- analytical methods, used for the prediction of motion of near-Earth natural or artificial celestial bodies. In order to guide and control the observing instrument it is enough to connect it to the computer and to `translate' the orbital elements given by the software into celestial co- ordinates.

The CIAX system especially CIAX-3 increased observation efficiency for Cassegrain test instruments at the early phase of Subaru telescope test observation. In order to control this system effectively and automatically, a control software for the entire system of the CIAX was developed. The software design goals are (1) redundancy for robust system, (2) the safety of the instrument by interlocking, (3) maximum efficiency by automatic control and (4) easy user interface for operator. In this paper, we describe the software which has been being tested through the telescope and instrument commissioning phase.

The 32-bit database applications with multiple document interface for Windows has been deployed to study fast variability of individual features in spectra that had been undertaken apparatus error correction (such as flat field correction). The channel scale of all spectra had been transferred to the wavelength scale (angstroms). The database application has been developed using client/server model. The TCP/IP protocol is used in client/server communication based on the Windows Sockets (version 1.1) functionality. The application supports Dynamic Data Exchange conversation in server mode.

The UV-Visual Echelle Spectrograph has been installed September 1999 at the Nasmyth focus of the second Unit of the European Southern Observatory Very Large Telescope, located at Mount Paranal in Chile. Commissioning has been successfully completed December 1999.

Adaptive Optics as a new tool for astronomical observation has proved a powerful means of investigation in high angular resolution programs. However, in spite of the complexity of the components involved (wavefront sensor, real-time computer), its use must be made as simple as possible in order to make it accessible to the largest audience of observers, and to answer the more demanding needs of modern observatories such as queue scheduling, service observing or remote observing. The Computer Aided Control developed for the Nasmyth Adaptive Optics System of the Very Large Telescope, will provide the astronomer with an extensive support, from the preparation of optimized observations to the automated operation of the instrument at the telescope either for hardware control, real time computing, or even preventive maintenance.

The object oriented software design paradigm has been instrumented in the development of the Adoptics software used in the Hooker telescope's ADOPT adaptive optics system. The software runs on a Pentium-class PC host and eight DSP processors connected to the host's motherboard bus. C++ classes were created to implement most of the host software's functionality, with the object oriented features of inheritance, encapsulation and abstraction being the most useful. Careful class design at the inception of the project allowed for the rapid addition of features without comprising the integrity of the software. Base class implementations include the DSP system, real-time graphical displays and opto-mechanical actuator control.

The Adoptics software used with the Hooker telescope's ADOPT adaptive optics system implements real-time control on a network of eight Texas Instruments C40 40Mhz DSP processors. System inputs and outputs consist of a 32 port 4 fiber wavefront sensor CCD with 32 X 32 pixels, a tilt mirror and a deformable mirror of which 241 elements are actively controlled. The network os eight DSP processors is configured as two cross-connected rings of four processors each. Four input sites perform intensity and gradient calculations immediately on incoming pixels as well as passing pixels on to adjacent sites for parallel calculations.

For CCD wavefront sensor pixel input and processing the Adoptics software in use at the Mt. Wilson 100-inch Hooker telescope utilizes hand-crafted assembly code to maximize use of the Texas Instruments C40 DSP processor's bandwidth. For flexibility the software is built around a conceptual framework that allows compact implementation of wavefront sensor-specific operations as well as adaptability to different CCD architectures. Designed for use in multi-DSP systems, such as those used at Mt. Wilson, the software framework will support many different DSP network topologies and facilitates distributing pixel-processing operations across the network.

New framework of data analysis system (DASH) has been developed for the SUBARU Telescope. It is designed using object-oriented methodology and adopted a restaurant model. DASH shares the load of CPU and I/O among distributed heterogeneous computers. The distributed object environment of the system is implemented with JAVA and CORBA. DASH has been evaluated by several prototypings. DASH2000 is the latest version, which will be released as the beta version of data analysis system for the SUBARU Telescope.

The new ESU Thermal Infrared Multi-Mode Instrument TIMMI2 is described in detail in Reimann et al. The TNT-Timmi Navigator Terminal is the graphical user interface for TIMMI2. It provides the communication between telescope, instrument and reduction pipeline. The TNT is a very easy way allows the astronomer to prepare and run complex observing programs. The graphical elements are based on Forms Library (A Graphical User Interface Toolkit for X). The TNT is written in C.

The Lowell Observatory Instrumentation System is the control system for a series of new instruments at Lowell, including the SOFIA first light instrument, HOPI. Sine these instruments will incorporate various detector systems and will be used with several telescopes, the concept of a loadable modulator based design was developed. The fundamental idea is to view the telescope, camera, and other instrument components as separate, interchangeable entities.