Timeline for Error probability for 8-PSK
Current License: CC BY-SA 4.0
15 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Feb 21, 2021 at 9:38 | comment | added | Marcus Müller | (By the way I'm very serious as ScienceDirect being really bad: I block them in my google searches, because I've never found any useful resource that way; trying to work through their misleading aggregation of publication excerpts has so far only been a time sink to me, and all the good things I've found were either in classical textbooks, or on Google scholar and Arxiv. Your problem is a case for a classical textbook, not sciencedirect, not Google scholar, and not arxiv) | |
| Feb 21, 2021 at 9:34 | comment | added | Marcus Müller | The wikipedia page: notice how they do exactly what I outlined above! First, they realize they need to derive a symbol error rate, which they call $P_S$. For that, they need a transformed probability magnitude function in angles, which they call $_\theta$, derived through a polar coordinate transform. They approximate the likelihood to use the $Q$ function, then they note that the BER can only be inferred from knowing the bitmapping! | |
| Feb 21, 2021 at 9:30 | comment | added | Marcus Müller | Could you also link me directly to the article, not to the sciencedirect "aggregation" page? Note that these pages are really bad as they simply do a text search and then present excerpts from hundreds of different publications of very different qualities and very different contexts, but without any context? You'll need to read at least the first two sections of a paper to decide whether it refers to your use case! | |
| Feb 21, 2021 at 9:27 | comment | added | Marcus Müller | Hi! Awesome, questions are always welcome :) So, I've been trying to go through your slideset, and I'm not sure which of the many examples you're referring to. Could you point me to a slide number? | |
| Feb 21, 2021 at 3:46 | comment | added | MagicMan | @MarcusMüller Regarding point three, and "random variable transformation PDF", do you mean this paper: math.arizona.edu/~jwatkins/f-transform.pdf ? Because I can't find that course. | |
| Feb 21, 2021 at 3:31 | comment | added | MagicMan | @MarcusMüller As for the issue of the formula and the source not being given - the formula, from what I've seen is some variation of the one on wikipedia (here: en.wikipedia.org/wiki/Phase-shift_keying#Bit_error_rate_2). I found it in this paper: sciencedirect.com/topics/engineering/m-ary-phase-shift-keying - From what I have read, for high values of Eb/N0 (such as in my case), you can use this formula | |
| Feb 21, 2021 at 2:44 | comment | added | MagicMan | @MarcusMüller wirelessinnovation.org/assets/Proceedings/2012Europe/… | |
| Feb 21, 2021 at 2:43 | comment | added | MagicMan | Hello @MarcusMüller, thank you for your extended response, however, I have a few questions and comments about it. I'll address point 1 first - unfortunately I find it challenging to define the noise model. The transcveiver example I'm using has no documentation (even though it's the official Packet-Based Digital Link from NI, duh [I'll put the link to the "presentation" below]). If it accomplishes anything (though I doubt it) - I know that I use the Galoises pseudonoise (PN) sequences generator in the transceiver to generate some bits. | |
| Feb 20, 2021 at 16:38 | comment | added | Marcus Müller | @DilipSarwate I hence focused on the last sentence of OP: "And if not, how can I calculate it differently?" and tried to explain how to calculate it differently, knowing full well that a 6-step overview is far from a "and that's the formula you drop in numbers", but given the lack of information, that formula is something nobody can give. | |
| Feb 20, 2021 at 16:36 | comment | added | Marcus Müller | @DilipSarwate I do see that it's a wall of text, and that writing that was my choice, not OP's, so I shouldn't be blamimg them for it. But what is the alternative. VTC "too broad"? VTC "unclear"? Answer "Q is not a parameter but a function, and $n$ needs to be defined by the context"? None of these answer OP's question. | |
| Feb 20, 2021 at 16:34 | comment | added | Marcus Müller | @DilipSarwate Hm, The question "what to use for Q" told me that the appropriate answer would include an introduction to why you need the Q-function (and giving the info that it's not a constant, but a function). "What to use for $n$?" is something I can't answer because that formula drops from nowhere and $n$ is undefined, which told me that the author's approach of "just copy paste a formula from a text without understanding the mathematical framework" is the problem, and not their ability to paramterize the correct formula. So, I tried to approach that! | |
| Feb 20, 2021 at 16:31 | comment | added | Dilip Sarwate | Marcus, the question asked what should be used for $n$ and $Q$ and your "wall of text" did not address the questions asked at all. Perhaps a better use of your time would have been to write a more succinct answer addressing what was asked instead of overwhelming the OP with details that are of no interest to him. | |
| Feb 20, 2021 at 13:27 | comment | added | Marcus Müller | Come on, this two page answer got accepted in less than 5 minutes. Please read and understand answers before accepting them. This makes me feel like you didn't actually want a true answer, but just someone to confirm that your approach is correct, and as you saw a complicated answer, you just gave up. That means I've wasted an hour of my life.... | |
| Feb 20, 2021 at 13:23 | vote | accept | MagicMan | ||
| Feb 20, 2021 at 13:34 | |||||
| Feb 20, 2021 at 13:20 | history | answered | Marcus Müller | CC BY-SA 4.0 |