Back to the Basics - Seti Signal Analysis

Message boards : SETI@home Science : Back to the Basics - Seti Signal Analysis
Message board moderation

To post messages, you must log in.

AuthorMessage
Profile Tom M
Volunteer tester

Send message
Joined: 28 Nov 02
Posts: 5121
Credit: 276,046,078
RAC: 462
Message 2033517 - Posted: 22 Feb 2020, 12:52:54 UTC

Or how we try to tell a signal from noise?

I have a friend who has a very good question.

At least from a Seti@Home point of view, how do we tell a Signal from "the Noise".

I am a layman (it took multiple tries to get through a Calculus I course).

So I am looking for a tutorial or series of tutorials that help me get my mind around "Doing Signal Analysis" with a bent towards Seti@Home type analysis.

I get that we need to "clean" the data of RFI from Earth resources.

But what are/is the basics of Signal Analysis? Yes, I will be re-reading the intro articles. But I want "more" :)

Thank you.

Tom
A proud member of the OFA (Old Farts Association).
ID: 2033517 · Report as offensive     Reply Quote
rob smith Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer moderator
Volunteer tester

Send message
Joined: 7 Mar 03
Posts: 20767
Credit: 416,307,556
RAC: 380
United Kingdom
Message 2033544 - Posted: 22 Feb 2020, 16:45:00 UTC

Assuming the the data is in the time domain (signal strength vs. time)
The most basic is the Mk1 eyeball - plot the data with time on the x-axis and signal on the y-axis - any really significant areas (noise or otherwise) will show up.

Next up is the shift the data from time base to frequency base, There are a number of ways of doing this, but the most common pair are Fourier Transform and Fast Fourier Transform, best are, as their names suggest related to each other, the first is a little more rigorous, but far slower on large data set. Then you get things like Gaussian Transforms, Wavelets, and many more. Some are better at one type of data than others, but are less universal that the Fast Fourier Transform - which, as its name suggests, is fast (relatively speaking), and there are hardware solutions available (RT-FFT, great fun but scores FF on the expense scale).

So once you have your data in the Frequency Domain (crudely speaking signal strength vs. frequency) you can do all sorts of analysis, such as peak FD-peak detection - at what frequency do the peaks occur, FD-mulitplet detection - is there a series of peaks at regular(ish) frequency interval (we've all seen discussion of doublets and triplets), harmonic analysis, and so on......

Each of these tools and techniques has had many books and papers published on them, and it is all too easy to find oneself heading into the depths of number theory as well as signal analysis. There used to be a "fairly simple" introductory book with a title like "Basic Signal Analysis" as many of the worked examples were based on paper and pencil approaches (long before desktop computers were even at the one-per-lab level, never mind the ubiquitous smart phone and laptop of today...). I daren't give you the titles of the books I have on the subject as they are aimed at the post-doc market, great for giving 1st/2nd year 1st degree students a nose bleed.

In some respects it is mostly the application of the outcome of calculus, so it is all about repeating a number of fairly simple equations and gradually homing in on the answer (and that's something computers are very good at - they don't get bored. Well HAL did, but that's a different story!)
Bob Smith
Member of Seti PIPPS (Pluto is a Planet Protest Society)
Somewhere in the (un)known Universe?
ID: 2033544 · Report as offensive     Reply Quote
Profile Tom M
Volunteer tester

Send message
Joined: 28 Nov 02
Posts: 5121
Credit: 276,046,078
RAC: 462
Message 2033576 - Posted: 22 Feb 2020, 22:10:35 UTC - in response to Message 2033544.  

===edit===
There used to be a "fairly simple" introductory book with a title like "Basic Signal Analysis" as many of the worked examples were based on paper and pencil approaches

===edit===
In some respects it is mostly the application of the outcome of calculus, so it is all about repeating a number of fairly simple equations and gradually homing in on the answer (and that's something computers are very good at - they don't get bored. Well HAL did, but that's a different story!)


Thank you Bob. I will do more "simple" reading as I try to understand what we are "doing".

I guess that is why the Seti@Home screen saver is so attractive. Right or Wrong it engages that MK-1 eyeball you were talking about. :)

Tom
A proud member of the OFA (Old Farts Association).
ID: 2033576 · Report as offensive     Reply Quote
Profile Tom M
Volunteer tester

Send message
Joined: 28 Nov 02
Posts: 5121
Credit: 276,046,078
RAC: 462
Message 2042759 - Posted: 3 Apr 2020, 13:11:36 UTC

ID: 2042759 · Report as offensive     Reply Quote

Message boards : SETI@home Science : Back to the Basics - Seti Signal Analysis


 
©2022 University of California
 
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.