(ORDO NEWS) — For more than 60 years, scientists have been looking in space for possible signs of a radio transmission that would indicate the existence of extraterrestrial intelligence (ETS).
During this time, technology and methods have improved significantly, but the biggest challenges remain. In addition to the fact that a radio signal of extraterrestrial origin has never been detected, there is a wide range of possible forms that such a transmission could take.
In short, SETI researchers have to guess what the signal will look like, but without any known examples. Recently, an international team led by the University of California at Berkeley and the SETI Institute has developed a new machine learning tool that models what a message from an extraterrestrial intelligence (ETI) might look like. It is known as Setigen, an open source library that could be a game changer for future SETI research.
The research team was led by Brian Brzycki, a graduate student in astronomy at the University of California, Berkeley. He was joined by Andrew Simion, director of the Berkeley SETI Research Center, and researchers from the SETI Institute, Breakthrough Listen, the Dunlap Institute for Astronomy and Astrophysics, the Institute for Space Science and Astronomy, the International Center for Radio Astronomy Research (ICRAR), and the Gergen Institute for Data Science.
Since the 1960s, the most common SETI method has been to search space for artificial radio signals. The first such experiment was the Ozma project (April – July 1960), led by the famous Cornell astrophysicist Frank Drake (the creator of the Drake equation). This study used the 25-meter antenna at the National Radio Astronomy Observatory in Green Bank, West Virginia to observe Epsilon Eridani and Tau Ceti at frequencies around 400 kHz in the 1.42 GHz band.
Since then, these searches have expanded to cover larger areas of the night sky, wider frequency ranges, and a greater variety of signals. As Brzycki explained:
“In the 1960s, the idea was to focus on the region around the well-known frequency at which neutral hydrogen emits radiation in interstellar space – 1.42 GHz. Since this natural radiation is distributed throughout the galaxy, the idea was that any intelligent civilization will be aware of it and possibly choose this frequency to transmit in order to maximize the probability of detection.
Since then, especially with the rapid development of technology, the SETI radio range has expanded in all areas of measurement.
“Now we can make measurements in the frequency range of several gigahertz instantly. As data storage systems improve, we can collect huge amounts of data, which allows for higher resolution observations in both time and frequency directions. surveys of nearby stars and other directions in the galaxy in order to capture as much as possible potentially interesting directions in the sky.”
Another important change was the introduction of machine learning algorithms designed to find transmissions in the radiophonic noise of space and correct radio frequency interference (RFI). The algorithms used in SETI research fall into two categories: those that measure voltage time series data and those that measure time frequency spectrogram data.
“The raw data collected by the radio antenna is voltage measurements; the radio wave induces a current in the antenna, which is read and written as voltage,” Brzycki says. “A radio telescope is really just an antenna, complete with a parabolic dish to focus a larger area of light, which increases resolution and brightness.
It turned out that the intensity is proportional to the square of the voltage. Next, we are interested in the intensity as a function of frequency and time (when and where the potential signal) “.
To get this, Brzycki says, astronomers start by using algorithms that compute the power of each observed frequency against time series input data.
In other words, the algorithm converts the radio signal data from a function of space and/or time to a function depending on the spatial frequency or time frequency – the so-called Fourier transform (FT). By squaring it, astronomers can measure the intensity of each frequency over the period of data collection.
“To get a complete spectrogram, an array of intensity as a function of time and frequency, we take a section of the voltage time series, get the FT, then repeat this process throughout the entire observation, so that we can effectively stack a number of FT data arrays on top of each other in the time direction “Brzycki added. “Once you decide on the temporal resolution, we calculate the required number of temporal samples and calculate the FT to see how much energy is in each frequency bin.”
The main search algorithm used by SETI researchers is known as the “incoherent DeDoppler tree” algorithm, which shifts the spectrum of radio waves to correct for frequency drift and maximize signal-to-noise ratio.
The most comprehensive SETI search program, Breakthrough Listen, uses an open source version of this algorithm, known as TurboSETI, which has provided the basis for many searches for “technosignatures” (aka signs of technological activity).
As Brzycki explained, this method has some disadvantages: “The algorithm makes the assumption that the potential SETI signal is continuous with a high duty cycle (meaning it is almost always “on”). Finding a continuous sinusoidal signal is a good first step,
Because TurboSETI targets straight-line signals that are always “on”, it may have a hard time capturing alternative morphologies such as wideband and pulsed signals. Additional algorithms are currently being developed to detect these other kinds of signals, but as always, our algorithms are only as effective as we think they are targeting.”
For SETI researchers, machine learning is a way to identify transmissions in raw RF data and classify numerous signal types. The main problem, Brzycki says, is that the astronomical community doesn’t have a dataset of ET signals, making supervised learning in the traditional sense difficult.
To this end, Brzycki and colleagues have developed an open-source Python-based library called Setigen that facilitates the creation of synthetic radio emission observations.
“Setigen facilitates the production of synthetic SETI signals that can be used in fully synthetic data or added on top of real observational data to create more realistic RFI noise and background,” Brzycki said. “In this way, we can create large datasets of synthetic signals to analyze the sensitivity of existing algorithms or as a basis for machine learning training.”
This library standardizes synthesis methods for analyzing search algorithms, especially for existing radio surveillance data products such as those used by Breakthrough Listen. “They come in spectrogram and complex voltage (time series) formats, so having a method to get simulated data can be really helpful for testing production code and developing new procedures,” Brzycki added.
Algorithms are currently being developed for multipath observations using Setigen to obtain simulated signals.
The library is also constantly updated and improved as SETI research develops. Brzycki and colleagues also hope to add support for wideband signal synthesis to aid search algorithms that target non-narrowband signals. More reliable SETI research will become possible in the near future when new generation radio telescopes become operational.
These include Breakthrough Listen, which will use data from the MeerKAT array in South Africa. There is also the Square Kilometer Array (SKA) project, a massive radio telescope that will combine data from observatories in South Africa and Australia.
These include MeerKAT and the Hydrogen Epoch of Reionization Array (HERA) in South Africa and the Australian SKA Pathfinder (ASKAP) and Murchison Widefield Array (MWA) in Australia.
Alas, the most limiting factor in SETI remains – our extremely limited coordinate system. When it comes down to it, astronomers have no idea what an extraterrestrial signal should look like because we’ve never seen one before.
This, paradoxically, makes it difficult to isolate technosignatures among the background noise of space. Thus, astronomers are forced to use the low-hanging fruit approach, which means looking for technological activity as we know it.
However, by setting parameters based on what is theoretically possible, scientists can narrow their search and increase the likelihood that they will someday find something. As Brzycki summarized:
“The only potential solution to this problem is some kind of unobserved machine learning study that minimizes our assumptions; work is underway in this direction. Setigen certainly relies on this assumption – the synthetic signals that can be obtained are heuristic in nature as the user decides what they should look like.
“At the end of the day, the library makes it possible to evaluate existing algorithms and create datasets of potential signals to develop new search methods, but the fundamental questions of where and when will always remain the best we can do, is to keep looking.”
At times like this, it’s helpful to remind yourself that the Fermi Paradox only needs to be solved once. Once we detect a radio transmission in space, we will know for sure that we are not alone in the universe, that intelligent life can and does exist outside of Earth and communicates with technologies that we can detect.
Contact us: [email protected]