Professor Radu! Thank you so much for your labour!! As an idea for next videos: imagine we have a rover on the snow and it can be rainy or snowy at the same time. How could we train our rover to understand where it is exactly? And adapt to any obstacles? For example, we need it to “fill in” all the selected square. We cannot rely on GPS - 4-6 meters accuracy is too much We cannot rely on wheels rotation as they can slip on the snow And moreover, we cannot rely even on LIDAR - as it might be not enough contrast and snow/rain might influence it so much. It seems to me it is an interesting problem to be solved Thank you!
Sounds like a very specific problem :-) If I were to handle something like this, where none of the sensors are good enough, I'd look for a way of combining them to maximize accuracy. If wheels sleeping is a real problem, consider the accelerometer as well. It doesn't 'slip' but drift is a real issue there, so need to keep it 'under control' using the GPS information. This is just speculation... what I would also get as soon as possible is data: someone controlling the rover in different conditions and use the user controls as ground truth for what I want the decisions to be.
Unfortunately, yes, I can't make more than one video per week :-) and after this course I think I'll switch to one every 2 weeks. It's been quite demanding... But hey! These lessons have a lot of things inside them, and playing with the playground you can learn many new things as well :-) should keep you busy for a while!
Here we call this algorithm: Mouse Labyrinth. 😂😂🐀🐀🐀 This is because it helps to find the shortest way as simply as possible. The funny thing is that depending on the application we can find this algorithm in searches within a binary tree 😁. But you were very creative in how to implement it in JS.👏👏😁👍
You're awesome, man! I am so impressed by your incredible coding skills and excellent explanations. Your work is truly top-notch. Thank you for sharing all this amazing knowledge with us. I just wish you could do more in other programming languages like C (raylib) or SDL2.0. best regards from Tokyo
Tokyo, wow. I was there once and even climbed Fuji-san. Sometimes I wish I do these things in something lower-level so I can have more efficient programs, but I want to keep the programming language light and... not bother with explaining how to set everything up :-\ for JavaScript people usually have everything they need already installed.
Dear Prof. Radu, Thank you very much for sharing your insights. I completely grasp your perspective, and it certainly gives me something to ponder for future learning opportunities. The idea of a specialized, paid course for advanced learners is particularly appealing, and I would eagerly enroll should it become available. In the meantime, please continue with your exceptional work. Your contributions are greatly valued. Should your travels ever bring you to Tokyo, it would be my utmost pleasure to host you for dinner. Rest assured, we know the perfect place for the best sushi in the city. Additionally, my family and I would be delighted to introduce you to some of Tokyo's most beautiful sights. Warm regards
Wow, thanks for the invite! Sushi is not really my thing, but maybe because I haven't been in the 'right place' :-D Let's see if I'll ever make a course like that. I do RUclips mostly for fun and... as a result teach whatever I'm in the mood for (not following some strict schedule / curriculum). I do listen to what people ask in the comments, though, and sometimes get motived by those as well :-)
The logic from the last lesson is the one that can be applied for multiple cars as well, all you need to do is add a sensor for cars in front and on the right, and add a node for it similar as the one for the traffic light. I don't teach this specifically in the course because experimenting with many moving things is not easy (too much change 😕), but that shouldn't stop you from trying.
Unfortunately, I can't :-( I've learned the concepts over 10 years ago and since then I've just studied research papers and seen stuff at conferences (I'm not really a 'book person').
I was... And I think I will switch to one video every 2 weeks after this course ends. I actually have content scheduled already until September, so now taking it easy to recover from a... difficult year, so to speak (many things going on + I was sick quite a lot).
Thanks professor Radu..
You're Amazing
Hope you find it useful! :-)
Professor Radu!
Thank you so much for your labour!!
As an idea for next videos: imagine we have a rover on the snow and it can be rainy or snowy at the same time. How could we train our rover to understand where it is exactly? And adapt to any obstacles?
For example, we need it to “fill in” all the selected square.
We cannot rely on GPS - 4-6 meters accuracy is too much
We cannot rely on wheels rotation as they can slip on the snow
And moreover, we cannot rely even on LIDAR - as it might be not enough contrast and snow/rain might influence it so much.
It seems to me it is an interesting problem to be solved
Thank you!
Sounds like a very specific problem :-)
If I were to handle something like this, where none of the sensors are good enough, I'd look for a way of combining them to maximize accuracy. If wheels sleeping is a real problem, consider the accelerometer as well. It doesn't 'slip' but drift is a real issue there, so need to keep it 'under control' using the GPS information.
This is just speculation... what I would also get as soon as possible is data: someone controlling the rover in different conditions and use the user controls as ground truth for what I want the decisions to be.
Greedy me looking forward to another video today. Guess it's going to be a long wait till next week Friday. Thanks Radu 💯
Unfortunately, yes, I can't make more than one video per week :-) and after this course I think I'll switch to one every 2 weeks. It's been quite demanding... But hey! These lessons have a lot of things inside them, and playing with the playground you can learn many new things as well :-) should keep you busy for a while!
@@Radu Well you should! Or else you won't pass next year! PS: Great job as usual Professor!
@@shaddow11ro :-))
Here we call this algorithm: Mouse Labyrinth. 😂😂🐀🐀🐀 This is because it helps to find the shortest way as simply as possible. The funny thing is that depending on the application we can find this algorithm in searches within a binary tree 😁. But you were very creative in how to implement it in JS.👏👏😁👍
I didn't know it has a name :-D
Wow what a wonderful surprise.... Thanks Radu.
My pleasure :-)
Great vid as usual.
Thank you!
admirable
Thanks :-)
You're awesome, man! I am so impressed by your incredible coding skills and
excellent explanations. Your work is truly top-notch. Thank you for sharing all
this amazing knowledge with us. I just wish you could do more in other
programming languages like C (raylib) or SDL2.0.
best regards from Tokyo
Tokyo, wow. I was there once and even climbed Fuji-san.
Sometimes I wish I do these things in something lower-level so I can have more efficient programs, but I want to keep the programming language light and... not bother with explaining how to set everything up :-\ for JavaScript people usually have everything they need already installed.
Dear Prof. Radu,
Thank you very much for sharing your insights. I completely grasp your perspective, and it certainly gives me something to ponder for future learning opportunities. The idea of a specialized, paid course for advanced learners is particularly appealing, and I would eagerly enroll should it become available.
In the meantime, please continue with your exceptional work. Your contributions are greatly valued. Should your travels ever bring you to Tokyo, it would be my utmost pleasure to host you for dinner. Rest assured, we know the perfect place for the best sushi in the city. Additionally, my family and I would be delighted to introduce you to some of Tokyo's most beautiful sights.
Warm regards
Wow, thanks for the invite!
Sushi is not really my thing, but maybe because I haven't been in the 'right place' :-D
Let's see if I'll ever make a course like that. I do RUclips mostly for fun and... as a result teach whatever I'm in the mood for (not following some strict schedule / curriculum). I do listen to what people ask in the comments, though, and sometimes get motived by those as well :-)
💖💖💖💖
:-)
Thanks. Waiting for more than one car example
The logic from the last lesson is the one that can be applied for multiple cars as well, all you need to do is add a sensor for cars in front and on the right, and add a node for it similar as the one for the traffic light. I don't teach this specifically in the course because experimenting with many moving things is not easy (too much change 😕), but that shouldn't stop you from trying.
Muito legal! Obrigado.
You're welcome :-) Thank you for watching!
can you recommend some books about AI for beginner i mean for people that have some program experience
Unfortunately, I can't :-(
I've learned the concepts over 10 years ago and since then I've just studied research papers and seen stuff at conferences (I'm not really a 'book person').
from where the (right_hand_rule.car) came from in saves how he did it ? im confused a bit, please help🙏
from where that json format text came?
From Lesson 3, at 31:10
ruclips.net/video/gqswXFkLXwc/видео.html
@@Radu thank you so much🙏
No problem!
Ai cumva si o varianta scrisa in Python ?
Am ceva, dar nu e asa avansata. E versiunea de la sfarsitul primului curs (cu masinile mergand in linie dreapta).
That load function could be shorter...
Which one? There are several 🙂
In any case, you're probably right...
@@RaduThe one you just coded, the one at 17:22
@@miniontoby ah ok... how would you improve it? Using Object.assign, maybe?
You look a bit tired and definitely need more relaxed schedule.
I was... And I think I will switch to one video every 2 weeks after this course ends. I actually have content scheduled already until September, so now taking it easy to recover from a... difficult year, so to speak (many things going on + I was sick quite a lot).
Take Care @@Radu