setiQuest Algorithms (Data/API's)
The setiQuest Data/API's project aims to answer an important technical question about the search for extraterrestrial intelligence (SETI): Assuming aliens are using technology to sending us radio wave signals, what kind of signal are they using and what is the best algorithm for discovering it? They might be sending intense pulses or continuous sine waves or something else; there are an infinite number of ways to send a signal.1 Recognizing that our small group of experienced scientists at the SETI Institute have built-in biases, we are looking for fresh ideas and efficient algorithms from all creative people. If you have ideas and the expertise to invent, implement and characterize would-be signal candidates, the we want your attention!
The first signal type and the algorithm which finds it are unknowable (until we find one). This doesn't stop imaginative people from trying to guess. Physical principles and present-day limits of technology help constrain the search. For example, x-rays take a lot of energy to create and require space-borne detectors to observe. Neutrinos have many good properties for SETI because they interact weakly with matter, but this same property makes them difficult to detect. Hence most SETI scientists focus their attention elsewhere. Such assumptions should be viewed skeptically, yet most SETI searches focus on less energetic parts of the electromagnetic spectrum (optical and radio waves).2 For radio SETI, the properties of interstellar space and earth's atmosphere push our search toward signal types that are not corrupted during transit between the stars (e.g. circularly polarized sine-wave). Also, Earth's culture has invented an extremely effective algorithm for detecting sine waves, the FFT (Fast Fourier Transform). For more information on the choice of sine waves and comparison with other signal types, see e.g. [SETI 2020, Harp, 2010].
Our goal is to find highly-efficient and plausible signal/algorithm pairs to augment the present search. Good algorithms cover a wide phase space of signal realizations with minimal computation time. The telescope output is a contiguous time series of digital samples of the electric field captured by the telescope: The measured electric field as a function of time. For conventional radio SETI, this time series undergoes an FFT (Fast Fourier Transform), which is then squared. The result is examined for "spikes" which represent sinusoids that are much stronger than the background noise. Less conventional approaches include: simple power thresholding (looking for pulses), time-variable sinusoidal signals (amplitude modulated, frequency modulated, phase modulated, polarization modulated, ...), autocorrelation, KLT, and matched filtering with "magic numbers " like pi, sqrt(2) or e.
We reach out to you, the technically competent citizen, engineer or scientist, and offer unfiltered telescope data for you to begin exploration. These raw data files contain information from a wide variety of targets. You can write your own programs to test your algorithms and search for signals, or you can join the setiCloud project and modify/use/create algorithms and test them on real data using the Amazon cloud.
We're looking forward to your participation in the SETI community at setiquest.org.
1 Because the electric field can take any value
2 Physicists often use "light" to mean EM radiation of any wavelength.