SERENDIP Takes a Great Leap Forward
Posted by Amir Alexander
26-05-2009 12:00 CDT
Just when SETI@home is celebrating its 10th anniversary, its older brother, Project SERENDIP, is getting a general makeover. In June of 2009 Dan Werthimer will lead a group of 8 students to the Arecibo Observatory in Puerto Rico to replace the aging SERENDIP IV system which has been idle for the past two years. Once the brand new state-of-the-art SERENDIP V is in place, the project will go back online and return to searching the skies for a signal from an alien civilization.
If SETI@home is now a respectable 10 years old, Project SERENDIP's roots go back much further. The first SERENDIP was built at U.C. Berkeley in 1979, and collected data from the radio telescope at the Hat creek Observatory in California. Since then the project's hardware has been upgraded repeatedly: In 1992 SERENDIP's 3d generation was installed at the Arecibo radio telescope, the largest and most sensitive in the world. Five years later, with help from The Planetary Society, this system was replaced at Arecibo by SERENDIP IV, which continued in operation for the next 8 years. And now SERENDIP V is set to take over and continue a long and persistent search that began three decades ago.
As SERENDIP was upgraded repeatedly over the years, each new generation represented a quantum leap over the capabilities of its predecessor. SERENDIP I, for example, could look at 100 channels at a time, which seemed pretty impressive back in 1979. But its successor, which began operations in 1986, could look at 65,000 channels at a time, and with SERENDIP IV the number was up to 168 million! Such exponential growth is unheard of in most scientific fields, but is not unusual in SETI, which is always on the lookout for new cutting-edge technologies for scanning the skies.
When Dan Werthimer and David Anderson launched SETI@home in 1999, SERENDIP had already been in operation for two decades. At Arecibo it had its own dedicated radio feed, used exclusively for SETI purposes. This arrangement suited SETI@home very well, and the new project took full advantage of the infrastructure that was already in place for SERENDIP. For several years the two projects shared the same radio feed and analyzed the same data from space – each in its own distinct way. Finally in 2005 they parted ways, when SETI@home was adapted to the new multi-beam ALFA receiver whereas SERENDIP continued to operate with its own dedicated radio feed.
SETI@home and SERENDIP are sibling SETI projects, run by the same team from U.C. Berkeley and sharing much of the same history. Conceptually however, they are very different – and complementary. Simply put, whereas SETI@home conducts an exquisitely sensitive search of a relatively narrow band of frequencies, SERENDIP conduct a rougher search of a much broader band.
As Werthimer likes to put it, every SETI search tries to optimize four factors: Sky coverage, spectrum coverage, sensitivity, and signal types. For three of these parameters, SETI@home is outstanding: the ALFA multi-beam receiver, from which it draws its data, conducts regular and complete surveys of the entire sky visible from Arecibo. Thanks to the unmatched computing power of its world-wide network of personal computers SETI@home is by far the most sensitive search ever attempted, and is capable of scanning the data for a broad range of signal-types. But the last parameter, spectrum coverage, is SETI@home's weak point: All the data, so exquisitely analyzed, is drawn from a band only 2.5 Megahertz wide around the hydrogen line of 1420 Megahertz. If the aliens are transmitting at any frequency outside this band, SETI@home, for all its power and sophistication, will never hear them.
Judged by the same standard, SERENDIP appears to be the reverse of SETI@home. The sky-coverage of the two searches is comparable, since in the early years of SETI@home they used the same receiver at Arecibo, and will do so again when the new version is installed in June. But the older project does not have the enormous computing resources of the SETI@home network, and therefore can never hope to equal its sensitivity or its range of signal types. What SERENDIP does offer is excellent spectrum coverage: SERENDIP IV could instantaneously cover a band of 100 Megahertz around the hydrogen, and with SERENDIP V this band will be increased to 300 Megahertz. That is 120 times wider than the band covered by SETI@home.
To accomplish this, SERENDIP operates very differently from its sibling project. It does not record all the data received in the manner of SETI@home, and then send it on for future analysis. Doing so for such a broad band would require recording hundreds of millions of data points each and every second, which is simply too much data for storage and systematic analysis. Instead, SERENDIP scans the signal from the radio telescope in real time, right when it comes down from the receiver. Most of the data SERENDIP lets pass unrecorded, and it is forever lost. But if the signal strength at a particular frequency stands out sufficiently above the background noise, the system records it and saves it for future analysis. How often this happens depends on the system's sensitivity threshold, which Werthimer and his team set in advance. In practice this comes to about a 100 recorded signals every second and while this may still sound impressive, it is a lot fewer than the hundreds of millions of data points that are discarded at the same time.
This, however, is only the beginning of the detection process. The tapes with the recorded signals are then packed up and sent to the SERENDIP team in Berkeley, with an additional copy going to colleagues at Cornell University in Ithaca, New York. Each group then uses complex algorithms to analyze the signals that stood out above the background noise. The algorithms take into account the strength of each signal, the location in the sky that it came from, and whether there is a star, or possibly a known planet, in that direction. Finally they look to see whether a signal had previously been received from the same direction, since a persistent signals is far more promising than a one-time flash in the pan. Any of these features, or a combination of them, could indicate that an innocent looking signal is in fact an intelligent transmission.
Like every previous generation of the project, SERENDIP V represents a giant step forward over its predecessors. First and foremost, the new system improves on what has always been the project's greatest strength – spectrum coverage. In the first stage, SERENDIP V will cover a band of 200 Megahertz on both sides of the hydrogen line – double the bandwidth of the older system. In reality the improvement is actually fourfold, because the new system will search for a signal in two polarizations, whereas the older system only searched one. At some point in the future the spectrum coverage will be increased to 300 Megahertz at two polarizations, which is six times the older capabilities.
The increased bandwidth will enable SERENDIP V to listen to the entire spectrum of the Arecibo receiver that will be collecting its data. This is the ALFA multi-beam receiver, the very same one used by SETI@home since 2006. ALFA is a modern receiver with excellent sensitivity and a broad spectrum, but it also has this unique feature: Unlike an ordinary receiver, which collects data from one point in the sky at a time, ALFA's seven beams point to seven locations simultaneously, and listening to all of them. This unique feature enables ALFA to cover the sky much faster and more efficiently than traditional receivers, and makes it and ideal instrument for the sky surveys conducted by SERENDIP and SETI@home.
SERENDIP and SETI@home are sibling projects, representing different but complementary approaches to the search for an alien signal. If, as many SETI researchers believe, a far-away civilization is broadcasting at the magical frequency of the hydrogen line, then their signal will most likely be picked by SETI@home. No other search can approach SETI@home's sensitivity, which is made possible by its global network of computers. But if the aliens, for their own reasons, are broadcasting at a different frequency, then their call is more likely to be heard by SERENDIP, listening to its entire broad band.
Which approach is the right one for SETI? Undoubtedly both: since we have no way of knowing in what form that elusive signal will come, SETI scientists agree that we should cover as many possibilities as possible. Which system is more likely to detect a true signal from aliens? Only time will tell.