Introduction
"In space, no one can hear you scream." Just ask Sigourney Weaver or any other Hollywood astronaut, who rely on their wits when in peril and alone in deep space, far from would-be rescuers.
Indeed, space is an inhospitable place for humans, assuming you can afford to get there in the first place. In the changing economic and political climate of the late 20th century, the space science community saw an increasing number of scientific goals being met through remote observation and the use of robotic spacecraft. Such missions could be conducted at a fraction of the cost and with limited danger to humans.
But space and distance can also be inhospitable to the machines and technology involved in remote observations and sensing. Sending humans to repair or maintain remote technology may be costly or impossible. Consider, for example, the Hubble Space Telescope, which was found to have faulty optics after being launched into orbit and tested in situ for the first time. A costly embarrassment to NASA, the HST only began returning clear images following a subsequent shuttle repair mission.
This paper will explore the advantages and disadvantages of remote observation using both ground and space-based instrumentation from the standpoint of their scientific contribution to astronomy. The paper will also speculate on the future of remote observation and its potential contribution to science.
Definitions
Using a very liberal definition, everything we know about the world around us is through remote observation, beginning with the use of our eyes and ears.
Photons reflected from the surface of an object enter the eye, striking the retina. The resulting electrical stimulus is transmitted along the optic nerves to the visual cortex of the brain where it is processed, resulting in the perception of sight.
That is, the following physiological components work together to create the perception of sight:
In a more general sense, remote observation is defined as the use of remote instrumentation to collect data that is transmitted along some communication pathway for local processing and interpretation, usually with a scientific objective.
Sometimes, the more specific term, remote sensing, is used. This usually refers to observations of the Earth or other bodies from orbiting artificial satellites. Applications include the identification and monitoring of natural resources and surveillance.
In this paper, the more general definition of remote observation will be used, and will include the use of both remote ground and space-based instruments. The discussion will be confined to applications with scientific objectives.
Requirements for remote observation
Similar in purpose to their physiological counterparts in the human visual system, remote observation systems usually include the following components:
History and motivation
When Galileo Galilei pointed his telescope towards Jupiter in 1609, he sketched what he saw through the eyepiece of his telescope for subsequent analysis and publication. These drawings recorded a series of observations from which there was an inescapable conclusion; there were at least four bodies in orbit about the planet Jupiter.
Similarly, in the mid 19th century, William Parsons, the Third Earl of Rosse, pointed the "Leviathan of Parsonstown" towards M51 and discovered its spiral structure which he meticulously recorded in hand sketched drawings.
The contribution of these pioneers of astronomy is indisputable despite being based on visual observations alone.
Their work stands in stark contrast to that of present day astronomers, who utilise modern photographic and electronic detectors to capture their data; a reality that is somewhat at odds with the popular romantic notion that astronomers actually peer through the eyepieces of their telescopes. In this sense, virtually all modern astronomical observations are "remote" since the initial inspection and interpretation of raw data is almost never done at the time of data acquisition.
In addition to the use of modern technology, the motivation for remote observation falls into a number of broad categories, each of which will be discussed in greater detail in the discussion which follows:
Utility- Remote observation utilising robotic telescopes and automated image processing algorithms are utilised by research projects that could not easily be conducted through manual means due to time constraints, geographical constraints, or other factors. Examples include automatic detection of supernovae. Such events are unpredictable and of limited duration. Consequently, they must be detected in real-time to maximise the scientific value of post-detection observations. Quick reaction to targets of opportunity is a major advantage of remote automated robotic observations.
Convenience- Most astronomical observations are of short duration and usually conducted over a period of several hours to a single night or two. Consequently, it is often uneconomical to expect astronomers to travel to remote locations to oversee short observing runs. Furthermore, it is often unnecessary to expect astronomers to endure the physical discomfort of long travel times and the effects of working at altitude or in other unusual environments. These considerations have led many professional observatories to develop the infrastructure to support remote observation. For example, the Keck telescope in Hawaii affords mainland astronomers with the ability to conduct remote observations without enduring the time and expense of a mountaintop Hawaiian odyssey, while still allowing them to work collaboratively in real-time with technicians located on-site. Similarly, observing from a remote location is also available to astronomers using the Multiple Mirror Telescope (MMT).
Necessity- While the Earth's protective atmosphere is transparent at optical wavelengths, it is relatively opaque in many other important wavelength ranges. ESA's Solar and Heliospheric Observatory (SOHO) have enabled the sun to be observed without obscuring atmospheric effects at many wavelengths. Even at optical wavelengths, the Earth's atmosphere is unsteady and results in blurry images when acquired using traditional ground-based instruments. While adaptive optics may improve upon this problem, placing instruments and their detectors in Earth orbit eliminates the problem altogether. Examples include the Hubble Space Telescope (HST).
Robotic Instruments
Ground-based robotic telescopes and remote observing platforms in orbit about the Earth and other bodies have been used for a number of purposes. These include:
A significant scientific application of robotic telescopes, however, is automated detection of transient targets of opportunity.
For example, the University of California has conducted automated supernovae searches in galaxies visible in the Northern Hemisphere using the Katzman Automatic Imaging Telescope (KAIT). In the five years following the commencement of this research program in 1995, KAIT has been responsible for the discovery of 100 supernovae.
In the Southern Hemisphere, the Perth Astronomy Research Group (PARG) also has an active research project dedicated to automated supernova detection. A typical robotic observatory, similar to that used by PARG, is shown in Figure 1 [Williams, 1997a]. In such a system, one or more computers is responsible for functions including opening the observatory dome and pointing the telescope, CCD camera, filter wheel and focusing control, and image processing.
In the PARG supernova detection system, images for a set of target galaxies are acquired and automatically compared to an older library image of the same galaxy. In an initial pre-processing step, each image undergoes bias subtraction and flat field correction to remove the zero signal and correct for non-uniform pixel sensitivity. The new image is then aligned with the library image of the same galaxy. "Star-like" objects are identified by considering a region centred on each pixel and determining if the intensity of the neighbouring pixels vary by more than a specified threshold value. Automated analysis follows to determine if star-like regions in newly acquired images are statistically different from the original library image of the same object.
The PARG library image for NGC 2442 is shown in Figure 2a. A supernova in this galaxy was detected by the PARG detection system in 1999 and designated SN1999ga. The detection image is shown in Figure 2b, with the supernova indicated by a red arrow.
While manual set-up is required to collect flat field images at the beginning of each observing session, it is reasonable to speculate that this part of the system could also be automated given additional resources.
The PARG system detected 14 supernovae between 1993 and 1999. In its initial three years of operation, a total of 5540 galaxy images were taken, utilising a total of 370 hours of telescope time, with most fields observed and re-observed at two week intervals [Williams 1997a, 1997b]. While manual inspection of images may have been possible given sufficient human resources, this operation is well suited to an automated approach that guarantees consistent and reliable application of the detection criteria.
Remote observation from artificial satellites and spacecraft
Target of opportunity observations are not limited to those made by ground-based robotic instruments.
During the Cold War, US spy satellites were sent into orbit to detect gamma radiation generated by enemy nuclear explosions.
A surprise at the time, gamma ray bursts were seen to occur with alarming regularity. Not produced by nuclear weapons, these bursts were apparently produced by naturally occurring phenomenon.
Unravelling the mystery of the gamma ray bursts, though, would require the cooperation of gamma and x-ray satellites with other optical and radio instruments both on the ground and in space.
When a gamma ray burst is detected, gamma and x-ray satellites pinpoint the burst's location in the sky and alert a network of other optical and radio instruments which attempt to identify the source and begin follow-up observations. Once successful, the transient optical and radio signature of the events were recognised; they were consistent with those produced by supernovae, the implosion of a massive star resulting in the creation of a black hole. What's more, it has been shown that bursts tend to occur in galaxies with a tremendous red shift, and hence are located at great distances and from the time of the early universe.
There is still much to learn about gamma ray bursts. However, without space-based robotic instruments to pinpoint the direction of gamma ray sources, other optical instruments both in space and on the ground would not have known in which direction to look for further analysis; the nature of the bursts would still be a mystery.
The Hubble Space Telescope (HST) is an important component of the network of instruments that have worked together towards unravelling the mystery of gamma ray bursts. But this is not the only contribution of the HST. Other significant contributions include the following [HREF 8]:
One of HST's crowning achievements, though, has been the so-called Hubble Deep Field Image. This observation has enabled astronomers to look further back into the distant universe than previously possible.
To produce the Deep Field Image, the HST was pointed to an apparently empty area of the sky for 10 days. The resulting image, a composite of many long exposures, reveal a multitude of galaxies, including many well-formed spiral and elliptical galaxies.
The Deep Field Image and its counterpart from skies of the southern hemisphere has led astronomers to the conclusion that galaxy formation was well under way as early as 1 billion years after the big bang. Unfortunately, galaxies in the early universe are red-shifted well into the near infrared range of the electromagnetic spectrum, so looking back to a time before that pictured in the Deep Field Images will require a new space-based telescope that is sensitive in that spectral region.
Future
The Next Generation Space Telescope (NGST) currently on the drawing boards of NASA at its contractors will be sensitive to infrared radiation, thereby allowing astronomers an unprecedented look at the early universe.
To achieve this goal, the instrument must operate at very cold temperatures, so current plans suggest that an NGST orbit of 3 Astronomical Units (AU) from the sun may be appropriate. To deliver a payload to this distance, the instrument must be small and folded to fit in the payload bay of existing delivery systems and deployed once in position.
Such a space-based system could be the forerunner of other highly sensitive instruments with high resolution and perhaps capable of directly imaging Earth-like extrasolar planets.
It has even been suggested that the proposed Terrestrial Planet Finder (TFP) mission may be capable of detecting the spectral signature of biological activity on such a world, if it exists, and take us closer to determining if humanity is alone in the universe.
Indeed, ordinary people around the world have indicated their interest in answering this question through their support of the SETI@home research project. In a certain sense, the over 2.75 million SETI@home participants are remote observers, twice removed. Their computers process data, collected using piggyback receivers eavesdropping on the observations of others at the Arecibo observatory, and shipped on tape to be re-distributed by servers at the University of California at Berkeley.
Conclusions
While existing space-based robotic observing platforms have increased our understanding of the universe in a manner that would be impossible from the ground, it is sobering to consider the lessons of the HST, which required a repair mission to fix faulty optics.
Orbiting far beyond the reach of the Space Shuttle and existing spacecraft capable of carrying human repair personnel, it will be important to get NGST and TFP right from the start. Repair missions to fix design faults or to include new instruments will likely not be possible. NASA will have to get it right the first time.
Assuming they do, humanity may soon have answers to age old questions. We will be closer to understanding the origins of the universe, and may even come to know if we are alone in the vastness of space.
References
APOD (HREF 1) The Hubble Deep Field, http://apod.gsfc.nasa.gov/apod/ap000709.html
APOD (HREF 2) The Hubble Deep Field South, http://apod.gsfc.nasa.gov/apod/ap981214.html
(HREF 3) Bradford Robotic Telescope, http://www.telescope.org/rti/use.html
(HREF 4) Katzman automatic imaging telescope, http://astron.berkeley.edu/~bait/kait.html
Kjeldseth-Moe, 0 (HREF 5) Solar instruments in space 1946-1999, http://www.astro.uio.no/~olavm/solar_space_instr/solar_space_instr.html
MMT (HREF 6) The MMT Observatory, http://sculptor.as.arizona.edu/foltz/www/mmt.html
MMT (HREF 7) Remote Observing at the MMT, http://sculptor.as.arizona.edu/foltz/www/remobs.html
NASA (HREF 8) NASA Facts: Hubble space telescope yields unprecedented scientific accomplishments in the first decade of observations, FS-2000-03-002-GSFC, http://www.gsfc.nasa.gov/gsfc/spacesci/hst10/HSTTop10.pdf
NASA (HREF 9) Discovery may be "Smoking Gun" in Gamma Ray Burst Mystery, http://science.nasa.gov/newhome/headlines/ast31mar97_1.htm
NASA (HREF 10) NGST: The Next Generation Space Telescope Executive Summary, http://www.ngst.nasa.gov/project/bin/BreckenridgeExecutiveSum.pdf
PARG (HREF 11) Perth Astronomy Research Group, http://www.parg.asn.au/
REACT (HREF 12) The REACT optical GRB follow-up network, http://pulsar.ucolick.org/REACT/
Shopbell, PL, Cohen, JG and Bergman, L (1998) Remote observing with the Keck Telescope: ATM networks and satellite systems, Astronomical data analysis software and systems VII, ASP Conference Series, Vol. 145, 1998, Ed: Albrecht, Hook and Bushouse, http://www.stsci.edu/stsci/meetings/adassVII/shopbellp.html
Williams, AJ (1997a) The Perth Automated
Supernova Search, PhD Thesis, University of Western Australia,
Abstract: http://www.physics.uwa.edu.au/~andrew/thesabs.html
Thesis: http://www.physics.uwa.edu.au/~andrew/MyThesis.pdf
Williams, AJ (1997b) Initial statistics from the Perth Automated Supernova Search, Publications Astronomical Society of Australia, vol. 14, no. 2, p. 208-13, http://www.atnf.csiro.au/pasa/14_2/williams/paper/
