![James Han](/img/default-banner.jpg)
- 14
- 105 461
James Han
Registrace 26. 01. 2013
Welcome welcome! I'm a first year robotics graduate student at the University of Toronto (U of T). I have an Engineering Science degree from U of T in Machine Intelligence.
This channel is literally just to have some random fun and to hopefully entertain and potentially teach some stuff on the way.
This channel is literally just to have some random fun and to hopefully entertain and potentially teach some stuff on the way.
The Generalized Gaussian Filter (Used in KF, EKF, and UKF)
The Generalized Gaussian Filter is not really a filter but a powerful tool for deriving other Gaussian filters like the KF, EKF, or UKF. The Generalized Gaussian Filter is a two-step process that relies on manipulating Gaussian probability distributions. Once it is derived, you can just use it as a look-up formula: if I have this, then I can write this. Super easy to use. Hope you enjoy!
Chapters:
0:00 Video Introduction
0:25 Overview of Process
2:11 Mathematical Derivation
3:24 Example 1: LG System
5:34 Example 2: UKF
Chapters:
0:00 Video Introduction
0:25 Overview of Process
2:11 Mathematical Derivation
3:24 Example 1: LG System
5:34 Example 2: UKF
zhlédnutí: 545
Video
30 Day Piano Challenge
zhlédnutí 149Před 10 měsíci
I haven't touched a piano in over 9 years... This challenge was a little jumpstart for me to start playing and more importantly, enjoying the piano again. Even with as little as 30 minutes of piano a day, some pretty satisfying progress can be made! Also if you were wondering... the Viva La Vida "Music Video" was made in less than 2 hours. Chapters: 0:00 Intro 0:30 Motivation for Challenge 1:31...
The Particle Filter: A Full Tutorial
zhlédnutí 5KPřed 10 měsíci
The Particle Filter is one of my FAVOURITE algorithms. It's so simple to understand and to implement, yet the performance is quite robust! The central idea behind the particle filter is to brute force your way to the solution. Start with a bunch of particles that represent where you think you are right now. Then, for the prediction step, propagate each particle through the motion model. Followi...
Importance Sampling: A Rigorous Tutorial (A Must-know for ML and Robotics)
zhlédnutí 1,1KPřed 11 měsíci
Importance sampling is a technique used when you have a probability distribution that is difficult to sample from. It uses a distribution we can actually sample from; we use those samples to estimate quantities regarding our other complex distribution. It almost seems like magic... But I guess sometimes math is magic :) Google Colab Notebook: - colab.research.google.com/drive/1qxlcQivhXjham2jZF...
How I Learned to Juggle and Became a Better Athlete in Just 2 Weeks
zhlédnutí 118Před rokem
Juggling! A talent for display? Or the secret to athletic success? Why not both? Learning to juggle has always been on my "to-learn" list, so when I started watching Drive to Survive (I highly recommend btw), my interest peaked when I noticed that a lot of drivers knew how to juggle. So, after digging around online, I found out juggling has massive benefits to both motor and mental abilities. I...
The Unscented Kalman Filter (UKF): A Full Tutorial. PS. Sampling Methods are Amazing
zhlédnutí 17KPřed rokem
The Unscented Kalman Filter (UKF) is considered the best Gaussian Filter in terms of performance. It relies on the unscented transform, a powerful tool for transforming distributions. The process involves intelligent sampling of points from the initial distribution, which are then passed through the true transform function. The resulting samples are recombined to estimate the final distribution...
How I trained for 30 days to DESTROY my mile time
zhlédnutí 986Před rokem
It's been a LONG time since I last ran; my flat feet have always caused too much pain to justify running, so I play/do other sports/activities that don't aggravate the issue. But since I recently obtained custom orthotics, I thought it would be fun to see how far I could push my running ability in 30 days using the mile as a benchmark. Hope you enjoy! :) #running #transformation #runningmotivation
The Extended Kalman Filter (EKF): Why Taylor Expansions are Awesome
zhlédnutí 7KPřed rokem
In this video, we explain how to derive the Extended Kalman Filter (EKF) from the definition of the Bayes Filter and how it is closely related to the Kalman Filter. The EKF is widely used because it is a simple yet effective way to handle nonlinearities in the motion and sensor models. This video will naturally lead into the Unscented Kalman Filter (UKF), which is another method to handle nonli...
The Kalman Filter Derived: The Power of Gaussians
zhlédnutí 2,5KPřed rokem
In this video, we derive the Kalman Filter from the definition of the Bayes Filter. The Kalman Filter is the simplest of the Gaussian Filters but it is crucial for the understanding of the more flexible (and incredibly popular) Extended Kalman Filter (EKF) and Unscented Kalman Filter (UKF). #robotics #tutorial
The Bayes Filter: A Tool Every Roboticist Should Know
zhlédnutí 6KPřed rokem
In this video, we provide a succinct overview of the Bayes Filter, its objectives, and the process behind its formula derivation. Serving as a foundation, this tutorial paves the way for exploring different Bayes Filter implementations, including the Kalman Filter and Particle Filters. #BayesFilter #robotics #tutorial
the problem with this explanation is that it's a mindless set of whats, eg: do this then do that then do this etc. What would be really useful is every so often discussing the need for that step, why (not what) it's there etc. atm this video just sounds like someone reading out of a book with fancy diagrams.
Thank you for teaching me this!
are The variables μt bar and Σt bar are the same as the mean mt and the covariance Ψt (where here they are considered predicted to get a predicted x)?
super neat and intuitive. Thanks!
The link of "Generalized Gaussian Filter" book is dead.
Apologies! Try this one: asrl.utias.utoronto.ca/~tdb/bib/barfoot_ser17.pdf (page 103 and pdf page 120) Here is a video explanation if you prefer: czcams.com/video/oPgfa6G2AxE/video.html&ab_channel=JamesHan
@@jameshan8 Thanks!
It is proper insanity how streamlined and efficient your explanations are. Thank you so much for these amazing resources!
Full Tutorial? Simulink UKF Implementation????? Regards!
How does it compare with the Quadratute KF?
Great video! Do you mind sharing what music you used?
Thanks! It's Dream Escape - The Tides
where can we get the article about table of parameters?
Here: ieeexplore.ieee.org/document/882463
Thanks a bunch @@jameshan8
How did you calculate the torque required for lifting load with threaded rod
video is awesome, but the music in the background is totally unnecessary ://
Can u explain about this mechanism.
Full tutorial about the Generalized Gaussian Filter was just released! czcams.com/video/oPgfa6G2AxE/video.html&ab_channel=JamesHan
Motor attachment and configuration?
Great video and excellent explanation. I only have one question regarding your last slide where you list the UKF Advantages. In point number 1 you mentioned that we don't need to know the non-linear models of the motion and/or sensor. However, in order to do the Unscented Transform for the sigma point, we need to know the non-linear models. That is the part I am confused. Thank you for your time.
Great question! I meant that we don’t need the analytical form of the models. If the models were a black box, we could still use the UKF since we only need to pass points into the model. For example if the motion model was modelled with a neural network or some lookup table from experimental results, the UKF would still work. Let me know if you still want further clarifications!
Hi James! Thank you so much for sharing this video. It is beneficial to my research. I'm wondering would you mind to also talk about the Practical Filter?
Hi Frank! You're welcome! Unfortunately I'm not familiar with the Practical Filter, so I'm unable to help you... Good luck with your research!
Wow
You not only explain the algorithm well but also prove it without skipping steps. And you do them in less than 10 mins. Bravo. If I had discovered you earlier, I would not have wasted hours and hours of my time on CZcams watching videos from people who haven’t fully captured the essence of these algorithms and don’t know how to teach. Thanks for your service.
Thank you for the very kind words! I'm glad you found it useful :)
Hello James, we have interest on developing this driving scissor lift. Do you have WECHAT or WHATSAPP to talk details?
Cool stuff😍
amamzing explanantion
Thank you! I really appreciate it :)
Does the word scent here mean that it has something to do withd armpit deorant ???
Great question! It does actually! Here's a quote from the author himself (ethw.org/First-Hand:The_Unscented_Transform): "Initially I only referred to it as the “new filter.” Needing a more specific name, people in my lab began referring to it as the “Uhlmann filter,” which obviously isn’t a name that I could use, so I had to come up with an official term. One evening everyone else in the lab was at the Royal Opera House, and as I was working I noticed someone’s deodorant on a desk. The word “unscented” caught my eye as the perfect technical term. At first people in the lab thought it was absurd-which is okay because absurdity is my guiding principle-and that it wouldn’t catch on. My claim was that people simply accept technical terms as technical terms: for example, does anyone think about why a tree is called a tree? Within a few months we had a speaker visit from another university who talked about his work with the “unscented filter.” Clearly he had never thought about the origin of the term. The cover of the issue of the March 2004 Proceedings we’re discussing right now has “Unscented” in large letters on the cover, which shows that it has been accepted as the technical term for that approach."
@@jameshan8 So inclusion , he just made that word up ? I do not see any mention about deorant in your answer
Hey@@tuongnguyen9391! The quote above includes the phrase: "deodorant on a desk"
@@jameshan8 oh now I see it thank you. But it seems like scientist are bad at naming their product which is why we need the sale team.
congratulations for the channel, I'm following it and learning a lot of great quality!
Thank you so much! That means a lot to me :) I'm glad they help you learn
One of the best explanations 💪 Great video, thanks!
Thank you! I really appreciate it :)
increase stability and put the slide slots and pin of table and foot in opposite directions.
Correction: @4:52 the variance should be the second central moment (not the raw moment that I put up). The 2nd central moment is E[(x-M1)(x-M1)^T]
Nice video! At 2:45, should the instances of p(y_k | ..., y_{1:k}) be p(y_k | ..., y_{1:k-1})?
Yes! Thank you! Great catch; I'll pin so others can see it too :)
Really cool editing and understandable examples :)
Im having voltage, current, acitve power, temperature, rpm of a BLDC motor as the known paramerters. Can i use kalman filter to estimate the torque with the help of the known parameters or is there any other simpler methods to calculate the torque
I wonder how did you attached the motor to the system and how did you figure out the appropriate screw
Excellent video, lots of good information in a very watchable format. Consider also covering the square root UKF, utilizing Cholesky decomposition, QR decomposition, and Cholesky rank 1 updates, it can perform significantly faster than UKF or even EKF while avoiding the UKF’s pesky issue of the covariance matrix losing positive definite-ness in the presence of poor/infrequent sensor updates.
Interesting! I didn't know that! I'll pin this so others can learn from it too :)
Full Video! czcams.com/video/h0C0C0B45rY/video.html&ab_channel=JamesHan
destroying mile times AND juggling???¿¿¿?? somebody stop this man
LOL as Hudson would say: TOO KIND. Actually my biggest supporter ❤
@@jameshan8 i support high quality content and high quality ppl >>:)
😍😍😍
ΚΑΛΑ, ΣΗΝΧΑΡΙΤΙΡΙΑ! 😄😄
IM LOVIN IT
Hi James. Awesome progress! Have you noticed an improvement with your shin splint pain? If so, what do you think helped the most?
Hey Ian! Thank you for the kind words and great question! I have noticed improvement but I don’t believe it’s from the orthotics. Running form is the biggest one for me! When I run slow and controlled in good form I never get shin splints, but when I speed up to medium pace I tend to heel strike more thus I almost always feel shin splints after roughly 600m. Also when I did get shin splints, massaging them tended to enable them to heal faster
Incredible quality! I was your 100th sub James!!
thanks Parker! And yay!! :)
How did you attach the motor
Hey! Excellent visuals and quality of video. Maybe the style of video isn't for me, but I certainly got lost (in this video and your later ones). I'm a stats / ML 3rd year student. Is the addition of additional information, the correction factor, or is it the cyclical fashion of the two parts conjoined, or is it the amount of belief we put in the sensor data? I did not have a clear grasp on what physical quantity corresponded to what abstract symbol in the derivation, and when we did get the derivation, I didn't understand what it showed us. I'm now really interested in the subject! There seems to be so much cool stuff here, the presentation just seems to have lost me. Will keep an eye on your channel for sure
Hey @mkwTL! Thanks for reaching out and providing some feedback :) I'm not 100% sure I understand the question but it seems like you're asking about what the sensor data is. The sensor data is often quantities like lidar measurements or distances to known objects. So @time=29s in the video, the yellow arrow is the sensor data. In localization, by knowing where the object is relative to me (in the localization problem we assume a known environment, but if we don't know the environment and need to estimate that as well then that is a SLAM problem which will come in a future video) I can generate an estimate of where I am. But in addition to just sensor information, if I know how I travel, I can use that information to determine where I am. So the idea behind the Bayes Filter is combining both of these types of information to generate a probabilistic estimate of where we are. I'm glad you're interested in this topic! Let me know if you have any questions or want to actually chat about it! My discord is Orblics#4757
Wow, really useful - thanks for putting this together! Clear derivation and great animations!
Cool project! I wonder why the deciscion for driving screw in the air instead of the bottom joints?
Can anyone tell me the formula needed to get the allowable speed at a specific height of the lifter for it to not buckle.
Hi. Would it be possible to contact you about this? Interested in your thoughts on building this for another application. Thank you.