Kalman Filter for Beginners

[@saadtiwana]

Thanks for explaining the intuition behind Kalman filter instead of just jumping into the mathematics right away. We need more videos like yours!

[@Augmented_AI]

Awesome. I'm glad you enjoyed it 😁. I focused on the most intuitive way to explain the concept

[@hanjarake_taro]

All the other guys except for you completely failed to explain what Kalman Filter is, which effectively means that they don't understand Kalman Filter. Therefore, you are the only one who understands Kalman Filter. Absolute god.

[@mayurnmahajan]

Finally understood the concept. Great teaching style.

[@erlfram]

Did NOT expect such good production quality from a video with 200 views and a channel with 500 subscribers. Very impressive!

[@Augmented_AI]

+erlfram you know comments like these make me smile :) . Thank you for the nice comment and I really appreciate the feedback. Really makes me want to give more value though my lectures :)

[@akshaynautiyal6644]

Thank you for this brilliant explanation !!!

[@lividpudding8565]

Augmented Startups thanks for the video!

[@AdamTheBot]

It's 135K views and 116K subs dude !! BYW are even alive ? If yes then please reply .

[@anuragthapa1395]

Excellent way of teaching. I got the gist of the Kalman filter finally.

[@Augmented_AI]

Glad it helped! Thank you :)

[@sureshkumar-cc1jq]

Great Job, you are the one who simplified the Kalman Filter explanation anybody can understand with simple fashion. You are a great teacher.

[@Augmented_AI]

Thank you Suresh, I am glad you feel that way and I am really glad that I can help make this easier to understand :)

[@PikachuCuteCat]

This is the best video ever made by anyone on anything.

[@juancuadra3697]

Your equation for position is missing a multiplication by "t" on your second term. It shall read Xf= Xi + Vi*t + 1/2(a)t^2, but this doesn't impact the foundation of your explanation. You may want to consider adding a note. Great Video! Thank you!

[@krishna96369]

If you want to get into specifics it should be "delta_t" not "t"😝

[@johne6081]

Very well done. Some of my grad. students want to use a Kalman filter in a vehicle line-tracking problem, and I would like to assign your video as an introduction to get everyone started on the concept.

[@Augmented_AI]

+John E Hi John thank you for the comment and you are more than welcome to show your students the video :). I am grateful to help. Do you have any other hard concepts that I can cover a video on?

[@fernandapaularocha7266]

Unscented Kalman Filter

[@ashishsheikh2151]

Particle Filters and Extended Kalman Filters ?

[@ghulamabbasawan7175]

i was looking to brush up my understanding of KF, which is about two decades old nows, and hence faded away. What a refresher, and such a wonderful and entertaining way of introducing a topic which is much involved. You are such a talented presenter. Please keep working on similar topics.

[@Augmented_AI]

Thank you. I'm really glad I could help 😁. Please shar with your friends

[@sherinkapoten]

Possibly my first comment on youtube in like years, only to complement on the teaching methodology. I learn't and remember less about the Kalman filter from my few years in grad than after having seeing this video!!!!

[@vudejavudeja]

Just about the right amount of information for me to get an idea of the concept. Well done!

[@Augmented_AI]

+vudejavudeja thank you I'm glad you enjoyed the video :)

[@pedrocolangelo5844]

Man, you're a genius. This explanation is incredible!

[@Augmented_AI]

Thanks you Pedro 😁

[@jp-hh9xq]

I have rewatched this video many times over the years. I use Kalman in my work in ADAS/AD, pretty much every day. I love this video. I want to make one for work. I actually showed your video in a group meeting once at a previous job and people had a hard time taking it seriously. That's on them. It is brilliant. I do want to steal some of your concepts though, to make a video I can show to my new group at work. I will credit you with the concept if I follow through. You nailed it though!

[@Augmented_AI]

Thank you JP. I'm glad you enjoyed it 😁. Yeah the key is to explain it to a 5th grader if it's on youtube. In a corporate setting you could swop out the examples for more relevant analogies. You may make your version and credit this video and channel, that would be great :)

[@ea1766]

By far the best video on this topic, just wow.

[@henryalferink1941]

EXCELLENT explanation! Thank you.

[@sridharsdrawingbook6316]

The best way of explaining Kalman. Thank you

[@Augmented_AI]

I'm really glad you enjoyed it 😊

[@evermelendez9732]

LMFAO!!!! "maybe the Pikachu slipped on a Rock!!!" by far the funniest, most engaging video I've seen looking for material on the Kalman Filter. Thank You

[@ThomasHaberkorn]

great video! please show something about using multiple sensors with the Kalman filter

[@vivek-159-icd]

A very good way of explaining the use of kalman filter

[@DXenakis]

at 0:52, you mention that now we want to estimate the initial position. The term "initial" is misleading.

[@bellicose2009]

it's the Initial position relative to the man. It's just an example to demonstrate the point.

[@tahaali3603]

Came to this video after searching for kalmans and listening to many others .. this made much ore sense and easier

[@Augmented_AI]

Im really glad you enjoyed it :D. Why dont you join our whatsapp group chat.whatsapp.com/JTuIB3eEfDRGo0TL4RzqwB

[@mattsmith2222]

Dude you got my full attention and I forgot to take my ADHD meds today. Killer video and super helpful :) Thanks!

[@FE_E_41_AdarshPandey]

Best vedio for learning kalman filter very good efforts

[@suheladesilva2933]

Really great video with an intuitive explanation. Thank you very much for your time in making this video.

[@srujanaturaga7382]

Didn't have a clue about it before watching this. GREAT example. Thanks alot! :)

[@Augmented_AI]

+srujana turaga thank you for the comment. I'm glad you enjoyed it :)

[@blacksamurai30]

I’m a surgeon working in a BMI lab, I’m not an engineer by any stretch of the imagination. This was the most amazing explanation of the filter ever.

[@Augmented_AI]

Haha I'm really glad I could make topic entertaining for you 😁

[@adelineng8255]

Kalman Filter concept explained simply. Easy to understand! Thank you!

[@yubrshen]

I really enjoy your teaching on 1-D Kalman filter. I hope that you can elaborate on 2D Kalman filter. I feel there is some complexity in how to combine two streams of measurements.

[@Augmented_AI]

+Yu Shen Hi yu. 2D is simple as 1D. You approach the problem as vectors.

[@muhammadabrarkhalid7426]

nice way to demonstrate tricky concepts. keep it up.

[@SachinNath-dj4lk]

Great video, continue the good work please.

[@SSJIV]

Excellent explanation. Thank you for your time and effort.

[@changtai02181]

Thank you, your graphical explanation is very clear, and it made me understand the concept.

[@Augmented_AI]

Im glad you enjoyed it Chang :). What would you like to see me cover next?

[@dhiroopulipaka7401]

best explanation of Kalman filter .. Thank you

[@gavinkistner772]

Fun presentation. I'm confused by the usage of EST(t-1) in "step 2". Are we sliding from the previous estimate to the new measurement, or are we LERPing between the current measurement and estimate?

[@Daxdax006]

love this, I'm recommending it to my class

[@Augmented_AI]

+John Ktejik thank you for your comment. I really appreciate it :)

[@PhoenixPerryisawesome]

Most awesome! I feel like I am being trained by a very experienced trainer! :D

[@jackbillings4109]

This was excellent. Please make more of these.

[@Augmented_AI]

+Jack Billings thank you Jack I really appreciate the feedback :D. Glad you enjoyed the video.

[@sudarsann5263]

Good way of teaching.. keep going

[@Augmented_AI]

Thanks a ton

[@LethalBB]

Didnt understand kalman at all until watching this. Great production.

[@Augmented_AI]

Hi Lethal, Thank you and I am glad you enjoyed this video :).

[@nyigambole1925]

"PDF.. not to be confused with adobe pdf" 🤣... you got me there... something that has been stuck in my mind since I first heard of probability distribution function.

[@Augmented_AI]

🤣🤣 funny story some people I've spoken to about this actually get confused about pdf and ask about Adobe pdf

[@gimbopgimchi]

Absolutely in love with your lecture lol.

[@Augmented_AI]

+Jin Kwon glad you love it :)

[@sagar11071994]

Great video! I was wondering if I can use your video to give a lecture on Kalman filters! I think it is a great way to create interest and make everyone learn/remember how KF works. Really well done!

[@thelikkhaparishe]

coolest tutorial ever!

[@Augmented_AI]

Thank you Lis 😆

[@kingshukbanerjee748]

Very nice introduction.

[@CLASHROYALE-sh2kb]

Thanks for the video, Super helpful to understand.

[@DeclanMBrennan]

1:40 Possible typo: Should that v in the first equation be v t ?

[@wexwexexort]

Great presentation.

[@haseebfaisal1806]

very good explanation. Easy to learn

[@Vic74297]

Really nice explaination !

[@MichaelGTadesse]

Thank you for the simplified explanation of Kalman Filter, I would appreciate it if you make another lecture on the use of Kalman Filter for Data Assimilation.

[@nickwinters2637]

At 6:02, there's the equation EST at t = EST at t-1 + KG(MEAS - EST at t-1). Is MEAS at t-1? or at t?

[@serdarbulut9087]

loved the dragon radar :D

[@practicalsoftwaremarcus]

The motion equation for Xf is lacking 't' term in the velocity 1:14 ! Awesome video !!

[@julianacienfuegos2370]

that equation of motion is missing a t. x = x0 + vt + 0.5at^2

[@yildizyeni]

exactly! I was just wondering how he added distance to velocity :)

[@Augmented_AI]

Thanks Julian, I have added an annotation to correct that.

[@mihir777]

Brilliant! Best intro on topic

[@Augmented_AI]

+Mihir Somalwar thank you so much :). I really appreciate it.

[@SarahReider]

Hello, my name is Sarah and I loved this video. My mom loved this video too. Now that she’s equipped with a massive understanding of Kalman filters, she can do anything. However I have a quick question - what happens if the Pikachu evolves into a Raichu? Does this change the optimal estimate?

[@Augmented_AI]

Haha then ash will need some mad skills to capture Raichu🤣. Glad you and your mom enjoyed the video

[@vinothbose]

Good Explanation. Thanks

[@Augmented_AI]

Glad you enjoyed it

[@imtiaznabi9411]

This is the sexiest explanation ever thank you

[@Augmented_AI]

Thank you 😎

[@wobby7055]

Haha so nice you picked up Pokemon as explanation context 👍👍

[@MEan0207]

How can I learn Full course the Kalman filter?

[@carlossantiago4845]

Excellent tutorial. I look forward to other videos like this.

[@songs1210]

AMAZING JOB!!!! LOVE THIS! People like you that will make our next generation geniuses.

[@Augmented_AI]

+Rich Francis thank you so much :). I really appreciate the comment. :) I'm glad to help 😊

[@Zakirkhan-nv3xw]

thank you. simple and concise explanation.

[@juansantiagocuadra3672]

I appreciate your time in creating such a useful content. Thank you.

[@akshaynautiyal6644]

Thank you for such a brilliant explanation !!!

[@Augmented_AI]

+akshay nautiyal thank you so much. It means a lot :)

[@samescobar7740]

this saved me, thanks

[@xstelaanax]

love it great job

[@iceymeng]

This video made my day! Thanks.

[@sushilkumarsingh3659]

Excellent video 👌

[@Sophie-fh3lh]

best video everrrrrr. tysmmmmm!

[@musicmoonshine]

Loving every second of this

[@Augmented_AI]

+James Almagest thank you, I really appreciate it :)

[@musicmoonshine]

+Arduino Startups if you ever pass through the south coast let me know :)

[@ayoogun5004]

Wonderful lecture! However, any work difference between particle filter and model predictive control?

[@RudolfEstragon]

Great video! I'm just wondering: we want to estimate where the Pikachu WILL be but we use the measurement of the radar at that point in time. So instead of predicting where it will be, we wait for the radar to measure it. Now it's more an improvement of the knowledge of the current position using our prediction and the measurement (which is also useful) but not really a prediction where to throw the ball. Am I getting this right or did I misunderstand something?

[@wendersonj]

Awesome explanation !!

[@Augmented_AI]

Thank you so much Wenderson :)

[@looper6394]

you did a common mistake at around 3:30. the y-axis is not the probability, its more like probability per meters and its highest value is not 1. in order to get the probability you have to integrate the pdf over an interval.

[@muhammadusama6040]

Wonderful explanation. Thank you very much.

[@Augmented_AI]

+Muhammad Usama you are most welcome :)

[@priyankajain-fb1bn]

how can i use this concept for self balancing robot? your help would be appreciated. thanks in advance.

[@ThomasHaberkorn]

there is a "t" missing in the first equation. it should read xf=xi+vi*t+1/2*at^2

[@eoril33]

very nice video, but I found the font quite hard to read!

[@rogeradi]

Great. Good introduction for me.

[@Augmented_AI]

+roger theyyunni thank you so much, I really appreciate it :)

[@Chuha97]

thank you for the simple explanation. The subtitle made by Anirban is terrible though...

[@PIYUSHTAILORstillalive]

3:24 didn't completed the video. But would say.. "Mazzzo aagyo, pura khel cover h bhaiiiii"

[@phillipneal8194]

Nice presentation. It would be great to see some python code for this. Thank you

[@mkuselimqana]

I like your style thank you

[@srinathbudhavaram5647]

many thanks for an excellent video on Kalman filter concept video. I know that, Extended Kalman filter is used for non-linear system state estimation. Could you please extend your video to cover Extended Kalman filter and Unscented Kalman Filter cases as well?

[@thaitoaninh1430]

good explain

[@Augmented_AI]

⭐ Haha, Thanks Thái Toàn Đinh, Also if you enjoy my work, Id really appreciate a Coffee😎☕ - augmentedstartups.info/BuyMeCoffee

[@saifghassan]

Great Job! Make a new one with EKF or SLAM

[@Augmented_AI]

+saif ghassan hey saif thanks for the feedback. I will definitely consider those topics :)

[@joelegger2570]

This video is really cool :) But the equation for X_f after 1:20 isn't correct. X_f = x_i + v_i * t + 1/2 * a * t^2 (You have to multiply v_i by t)

[@societyofrobots]

Is it fair to say that the Kalman filter is just a weighted average between the measured location (zero drift but low precision) and estimated position (high drift but high precision)?

[@madhaven694]

A big fat NO!! Do you see anywhere in the video, the estimate been divided by a number(constant)

[@ShahadatZ]

Thanks for this!

[@mohman222]

great video!

[@fabricejumel4630]

Very good !!!!!!!!!!!!! Congrats

[@Augmented_AI]

+Fabrice JUMEL thank you so much Fabrice glad you enjoyed it :) . really appreciate the feedback

[@tahaali3603]

Awesome best video

[@ACY55]

man you rock!

[@jonathan-bidwell]

Wonderful!

[@nicholaszammit8273]

Excellent

[@Augmented_AI]

+Nicholas Zammit thank you :).

[@phucthinhnguyen5633]

OMG!! I laugh to dead when the charmander appear.

[@ramradhakrishnan9382]

Looks like a great presentation, But.... Unfortunately, the audio is seriously muffled - I will have to build tune-able high pass filter with variable center frequency, bandwidth, roll-off, to process the audio from this presentation before I can follow it!.

[@ibrahimadeoti4430]

Thanks.