What is Autocorrelation?

[@payman_azari]

"Average Shared Directional Power" is this night's take to bed message; thanks for your amazing explanations professor 🙏

[@iain_explains]

Nice one! I'm glad you like the video. 😁

[@typicalidealist9138]

bro said "🤓"

[@agstechnicalsupport]

Autocorrelation well explained. Brief and to the point. Thank you for posting.

[@iain_explains]

Glad it was helpful!

[@beckyb8929]

speak for yourself. Clear as mud to me

[@osamahabdullah3715]

what a wonderful explanation, thanks a lot

[@iain_explains]

Glad you liked it!

[@kabuda1949]

Thanks for the explanation

[@iain_explains]

You're welcome

[@nataliexu8392]

Hey Prof Iain, so can i understand autocorrelation as a way to show the power information of a signal instead of really a correlation?

[@tuongnguyen9391]

Is there any digital communication scheme that actually use this concept ?

[@iain_explains]

It's used in the design of pseudo random sequences, that are used in CDMA and other low-probability-of-intercept signalling schemes, plus many other applications. I should make a video about this.

[@schlast8311]

well. I am here to try to learn autoorr([1 2 3 4 5]). I still have no idea what it is supposed to be.

[@ZiglioUK]

Autocorrelation can be indeed used to discover signals that exhibit pseudo-random noise characteristics, like CDMA

[@iain_explains]

Yes, that's right. Maybe I should make a video on that.

[@ZiglioUK]

@@iain_explains I saw first Balint Seeber suggesting using the auto-correlation to discover periodic PRN signals such as GPS. Then I saw a video tutorial in gnu-radio by jmfriedt where he performs a coarse GPS acquisition by squaring the signal. It's very cheap and convenient but there's a catch: 1. we see a signal but we don't know which satellite that is 2. the carrier modulation due to doppler doubles in frequency

[@arjunsnair4986]

Thank you sir

[@iain_explains]

You're welcome.

[@Will-bx2wz]

can you explain why we need to take complex "conjugate" E[Xt1, Xt2*] to calculate correlation? why won't E[Xt1,Xt2] (without conjugation) work?

[@iain_explains]

If you define it with the complex conjugate, then it means that for stationary processes, R_X(0) is real-valued, since X_t1 times X*_t1 is real (for any t1). And this is the "power" in the signal, since R_X(0) = E[ real(X_t1)^2 + imag(X_t1)^2 ]

[@iwillshinelikesun]

Without conjugation would work for real signals only. As x(t) = x*(t) for real signals.

[@rnpradios3938]

Your videos are great! Can please make a video on zadouff code

[@iain_explains]

Thanks for the suggestion. I've added it to my "to do" list.

[@ANJA-mj1to]

So valuable content that is superficially similar to the Convolution Theorem but as you present with differnt interpretation. For me this is can be so creating sample of water storage as civil engineer research. First of all thank you for 3 samples with important concept. All I see is full of knowledge and it can be so helpful - how you present type of signal in another time in 3 graphs; also how you present squere voltage with R>0 in another direction and imply in time average. Brilliant solved problem in theme summer and winter. We are concerned mostly with functions and manipulations like this to solve problem, but easy algebra is the key of success. This example will clearly show first path of building construction of the reservoirs downstream and upstream intended to isolate the water zones of the reservoir.

[@iain_explains]

I'm glad you liked the video.

[@greg77389]

"Voltage squared is power" Ummm, actually it's V^2/R 🤓

[@iain_explains]

In signals analysis we always assume a “normalised” unit resistor.