Gamma-ray bursts are violent explosions releasing invisible high-energy photons (gamma rays) and lasting only seconds to minutes at most. They are followed by a steadily fading afterglow emission of lower-energy photons, ranging from x-rays down through visible light to infrared and radio waves (see, "Gamma-Ray Bursts"). The afterglow typically lasts for hours to days.
Following its own logic, RAPTOR recorded the light signal every 30 seconds and noted a doubling in brightness over 4 minutes—an afterglow that was rising rather than fading. Running real-time analysis software, RAPTOR decided to report the anomaly to a human.
After receiving RAPTOR's call, Wozniak checked online and saw what looked like an explosive event. He quickly alerted Tom Vestrand, RAPTOR's team leader. Vestrand recalls his excitement, "This was a first, an autonomous optical telescope finding an anomaly on its own with no human intervention. If humans had been in the loop they would have said, as we did, ‘Gamma-ray bursts don't act like that. Forget it.' And RAPTOR wouldn't have found anything."
RAPTOR's observation of that spectacular "rebrightening" hints that a gamma-ray burst can sometimes "turn itself on" a second time, emitting intense visible light but no gamma rays—an intriguing possibility.
For the RAPTOR team, the discovery had a broader significance. It was proof that RAPTOR has a mind of its own—truly a thinking telescope system. This network of relatively small but rapidly aimed telescopes, coupled to a central computer, is able to define its own observing strategy and make discoveries in real time with no help. R2-D2, step aside!
The Los Alamos team is now trying to build a global network of upgraded fast-response telescopes that are hot-wired to a next-generation information system. Just as Google has revolutionized how we learn, the new thinking telescope network should serve as a "discovery engine" for astronomy, scanning the entire night sky every 5 minutes, screening a hundred million visible objects for ones with time-varying signals—optical transients—and picking out, on its own, the ones that have something new to tell us. It will then implement an effective observational strategy. Very ambitious, but what will we learn?
There are many kinds of optical transients, not just those from gamma-ray bursts, and they tell us about the dynamic evolution of the universe. For example, the light pulses from distant supernova explosions suggest that a radically new force—dubbed dark energy—is causing the universe's overall expansion to speed up, not slow down. The optical signals and afterglows of gamma-ray bursts are giving us hints about the various progenitors of stellar-size black holes. Because these bursts are found in the very-early universe, their bright afterglows may also teach us about the environments in which the first stars were born.
There's a thirst to learn more about dark energy and dark matter and about the unknown "life-cycles" of bright energy and bright matter—all the visible entities in the night sky, which are made of the same chemical elements we see on Earth. In 2012 the National Science Foundation, NASA, and the Department of Energy plan to initiate a very deep all-sky optical survey using the Large Synoptic Survey Telescope (LSST), an 8.4-meter telescope with a billion-pixel digital camera that will see halfway across the universe and pick up tens of thousands of optical transients every night.
But the LSST by itself won't be able to differentiate the important transients from the less-important ones. "The LSST could drown us in a flood of data," says Wozniak, "unless we have autonomous systems in place that are able to recognize the interesting events and follow up with real-time observations. Those systems will have to learn and evolve over time if we are to make sense of the dynamic database that is the night sky."
RAPTOR, with its impressive record observing the optical transients from gamma-ray bursts, the largest and most violent of stellar explosions, provides a solid platform for building a prototype system that can cope with LSST's torrent of data.
Gamma-Ray Bursts
Cosmic gamma-ray bursts, pulses of gamma rays, were detected in 1967 by the Vela satellites (sent into orbit to monitor the 1963 U.S.-Soviet Union treaty banning nuclear weapon tests in the atmosphere). The bursts' extreme energy output has continued to puzzle astrophysicists, but a general picture has emerged.
Rotating matter, in the form of a massive star or a pair of compact objects (such as two neutron stars or a neutron star and black hole) can reach a stage at which the force of its own gravity causes it to spiral inward. Because the matter is rotating, it also experiences a centrifugal force that opposes the inward motion. Gravity usually wins out, and most of the matter collapses to a point, forming a black hole—a region of space whose gravitational force is so strong that nothing can escape it, not even light. As matter plummets to the point of no return, it releases a huge amount of gravitational energy.
Now the picture gets fuzzier. In some unknown fashion, a fraction of that energy gets trapped in a disk of matter that was held back by centrifugal force and now rotates around the black hole.
The matter in the disk reaches velocities close to the speed of light and turns into a fireball consisting of electrons, positrons, protons, gamma rays, and probably magnetic fields. The fireball expands outward, forming a jet. As this jet of radiating matter and magnetic energy shoots into space at nearly the speed of light, it sweeps up whatever matter is in its path, producing a blast wave of gamma rays lasting tens of seconds on average. In its wake, the jet leaves hot, glowing material that produces a so-called afterglow (photons of much lower energy than gamma rays) lasting hours or days. The afterglow supplies information concerning the environment around the source of the gamma burst and further constrains how the burst might be generated.
When Los Alamos scientists began building robotic telescopes 10 years ago, gamma-ray bursts were making headlines because their location in the universe was finally becoming known.
These explosions, known since 1967, are so bright that the laws of physics seemed to demand that they originate in our galaxy. However, telescopes can't focus gamma rays, so it was impossible to trace them back to their source.
Then in 1997, a NASA satellite finally detected both a burst and the x-ray portion of its afterglow, which could be traced to a tiny patch of the sky. The world's largest optical telescopes—the Hubble in space and the ground-based Kecks in Hawaii—collected light from that patch for 4 to 6 hours and finally detected an ancient galaxy at the edge of the universe. Evidently the burst had come not from our Milky Way but from a galaxy billions of light-years away. One of the great scientific debates of our time was settled.
To be detected from so great a distance meant that the source of the gamma rays was astonishingly bright, releasing in tens of seconds as much energy as the sun would radiate in light in 10 billion years. Only the birth of a black hole could release that much power.
The race was on to see more of these bursts and to follow up each NASA satellite gamma-ray alert with studies of lower-energy signals (x-ray, optical, infrared, radio) that might hint at the source, or "central engine," of the gigantic energy release (see "Gamma-Ray Bursts").
Because these short-lived explosions occur only three or four times a day at random places in the universe, it seemed that optical telescopes, which must wait for satellites to spot the bursts, would never be able to follow up quickly enough to discover if an optical signal accompanied the burst itself. They would see only the afterglow that followed. Then, in a series of tour-de-force measurements, Los Alamos robotic telescopes did the impossible.
A Burst Caught in the Act
On January 23, 1999, RAPTOR's predecessor, ROTSE-I (Robotic Optical Transient Search Experiment-I), responded to a NASA gamma-ray burst alert and within 10 seconds recorded the very-early afterglow, the most-luminous optical source detected up to that time but not part of the burst itself.
RAPTOR images of the time-varying sky are the latest in "cosmic cinematography." This single frame from a several-hundred-frame movie in SkyDOT (Sky Database for Objects in the Time-Domain) shows the positions of a half-million objects, 7,000 of which are newly discovered variable stars. The next-generation RAPTOR will add to the movie at the rate of 100 frames per night.
Determined to catch a burst in the act, Vestrand and team improved and transformed ROTSE-I into the first generation of RAPTOR. Modeled on human vision, it had four small lenses that acted like the eye's peripheral vision and a central telephoto lens that captured the fine details.
Two of these systems placed 38 miles apart saw a patch of sky from two slightly different viewing angles (stereovision). A central command computer (RAPTOR's brain) continuously compared the two views to distinguish distant from nearby (mostly man-made) astrophysical objects and automatically remove thousands of small data glitches such as equipment noise and cosmic-ray hits.
Vestrand comments, "This distributed and yet integrated operation of robotic telescopes and data systems was the heart of RAPTOR and the ‘thinking telescopes' concept from the very beginning."
By December 19, 2004, a second-generation RAPTOR, with a larger 0.4-meter telescope and larger high-speed mount, was able to respond in only 8 seconds to a gamma-ray burst alert from Swift. The NASA satellite had seen a small gamma-ray pulse that turned out to precede the main burst by several minutes. RAPTOR was in position to record the optical emission throughout the burst and beyond. The recording revealed a complete surprise. Optical pulses were occurring during the burst, synchronized with the gamma-ray pulses. They must have been generated by the same "engine" that produced the burst itself. It was the first-ever recording of what is now called "prompt" optical emission.
On August 20, 2005, RAPTOR recorded another long burst that confirmed the previous findings. It also allowed scientists to say with a great degree of confidence that gamma-ray bursts have two distinct types of optical emissions contemporaneous with the gamma rays: the newly discovered prompt optical emission and the very-early afterglow, which starts up almost simultaneously. It's like a sudden energy release from an explosive charge, accompanied by embers glowing in the path of the blast wave.
Significance of the Optical Signals
RAPTOR's discoveries have set the stage for us to learn more about the central engine driving gamma-ray bursts and the medium immediately surrounding them.
The brightness of the optical signals suggests that RAPTOR could search independently for gamma-ray bursts without waiting for a satellite's gamma-ray alert. The afterglow's dramatic rebrightening that RAPTOR reported to Wozniak suggests that even more "optically rich," high-energy phenomena may exist and that they can be discovered only by surveying the sky for visible light. Each transient astrophysical phenomenon represents an area of intense interest that scientists want to study in depth.
But is it really possible to detect these short cosmic explosions from their optical signals alone, considering the complexity of the night sky? The gamma-ray sky is almost empty but for the few lone gamma-ray bursts and several hundred very faint persistent sources. By comparison, the visible sky is a maelstrom of flickering objects—satellites, airplanes, space debris, cosmic rays, meteors, asteroids, comets, flaring stars, active galactic nuclei, exploding novae, supernovae, and the rare optical transients from gamma-ray bursts—observed against a static backdrop of billions of ordinary stars and galaxies.
"The next version of RAPTOR will be a powerful new tool for searching out previously unknown phenomena within the maelstrom," comments Vestrand.
"Thinking" Goes Worldwide
RAPTOR's next incarnation will collect and analyze the plethora of optical transients at unprecedented rates (see "The Next-Generation RAPTOR"). A wide-field system such as the 16-lens RAPTOR-K will "harvest" time histories of all objects up to 10,000 times fainter than those detectable with an unaided eye. Tens of thousands of images collected every night, roughly 1 terabyte (a trillion units of data) per week, will be processed within seconds of collection and then compared with an enormous body of stored data on the time history of the night sky.
Emerging transient objects will be reported automatically while they are still occurring, giving response instruments enough time to make otherwise- impossible observations. In addition, the next-generation system will use arrays like RAPTOR-T (T for technicolor), carrying four co-aligned 0.4-meter telescopes with four different color filters, to pinpoint changes in the intensity of each color emitted during the first critical minutes of a gamma-ray burst. Information about the distance to the burst and about the burst's environment and dynamics may be gleaned from the color changes.
The Next Generation RAPTOR—A Closed-Loop System for Astronomical Discovery The next-generation RAPTOR will have smart, rapid-response telescope arrays at several locations around the globe. The on-site computers will identify interesting transients and send information about them to a central command computer that compares and interprets the data from the various sites. The command computer will also dispatch requests for information about local conditions affecting image quality, direct a coordinated observing strategy among all assets, and share information with satellites and ground-based telescopes outside the RAPTOR system, as well as with human overseers. The entire system will be continuously kept abreast of past and current developments through a centralized database of context data (images, observational conditions, and long-term time histories). The hierarchical structure will ensure coordinated real-time processing and rapid response to both internal and external alerts.
The speed of the system will continue to set records, but the truly revolutionary aspect is the introduction of "thinking" software agents in the system's command computer. These agents use modern statistical decision theory and algorithms that allow machines to learn in order to perform increasingly complex classification and anomaly-detection tasks, to carry on two-way conversations between the central decision-making computer and deployed assets about the state of the telescopes and the quality of the observed data, and to direct increasingly flexible allocation of telescope resources—all with no human intervention.
The RAPTOR team, riding on a wave of success, is now leading the development of an autonomous global network—a "discovery engine"—that can find totally new phenomena in the many uncharted regions of our time-varying universe. The team is hoping to be awakened many times in the coming years by phone calls from RAPTOR.
New-generation telescopes RAPTOR-K and RAPTOR-T, stationed on Fenton Hill. Like their predecessors, they perform their own routine housekeeping functions, including closing up during bad weather and alerting the central computer to equipment malfunctions.
1663
