The Fastest Maze-Solving Competition On Earth
소스 코드
- 게시일 2024. 04. 27.
- Welcome to Micromouse, the fastest maze-solving competition on Earth. Join Onshape’s community of over 3 million CAD users by creating a free account here: Onshape.pro/Veritasium.
Special thanks to our Patreon supporters! Join this community to help us keep our videos free, forever:
ve42.co/PatreonDEB
▀▀▀
A huge thank you to Peter Harrison for all of his help introducing us to the world of Micromouse - check out ukmars.org & micromouseonline.com.
Thank you to David Otten, APEC, and the All-Japan Micromouse Competition for having us.
Thank you to Juing-Hei ( / @suhu9379 ) & Derek Hall ( / @micromouse ) for usage of their micromouse videos.
Thank you to John McBride, Yusaku Kanagawa, and Katie Barnshaw for their help with Japanese translations.
▀▀▀
References:
Claude Shannon Demonstrates Machine Learning, AT&T Tech Channel Archive - ve42.co/ClaudeShannon
Mighty mouse, MIT News Magazine - ve42.co/MightyMouse
History, Micromouse Online Blog - ve42.co/MMHistory
Christiansen, D. (1977). Spectral lines: Announcing the Amazing Micro-Mouse Maze Contest. IEEE Spectrum, vol. 14, no. 5, pp. 27-27 - ve42.co/Christiansen1977
Allan, R. (1979). Microprocessors: The amazing micromice: See how they won: Probing the innards of the smartest and fastest entries in the Amazing Micro-Mouse Maze Contest. IEEE Spectrum, vol. 16, no. 9, pp. 62-65, - ve42.co/Allan1979
1977-79 - “MOONLIGHT SPECIAL” Battelle Inst. (American), CyberNetic Zoo - ve42.co/MoonlightSpecial
Christiansen, D. (2014). The Amazing MicroMouse Roars On. Spectral Lines - ve42.co/Christiansen2014
1986 - MicroMouse history, competition & how it got started in the USA, via KRplus - ve42.co/MMArchiveYT
The first World Micromouse Contest in Tsubuka, Japan, August 1985 [1/2] by TKsTclip via KRplus - ve42.co/MMTsukubaYT
IEEE. (2018). Micromouse Competition Rules - ve42.co/IEEERules
Tondra, D. (2004). The Inception of Chedda: A detailed design and analysis of micromouse. University of Nevada - ve42.co/Tondra2004
Braunl, T. (1999). Research relevance of mobile robot competitions. IEEE Robotics & Automation Magazine, vol. 6, no. 4, pp. 32-37 - ve42.co/Braunl1999
All Japan Micromouse 2017 by Peter Harrison, Micromouse Online - ve42.co/RedComet
Winning record of the national competition micromouse (half size) competition. mm3sakusya @ wiki (Google translated from Japanese) - ve42.co/JapanFinishTimes
The Fosbury Flop-A Game-Changing Technique, Smithsonian Magazine - ve42.co/FosburyFlop
Gold medal winning heights in the Men's and Women's high jump at the Summer Olympics from 1896 to 2020, Statistica - ve42.co/HighJump
Zhang, H., Wang, Y., Wang, Y., & Soon, P. L. (2016). Design and realization of two-wheel micro-mouse diagonal dashing. Journal of Intelligent & Fuzzy Systems, 31(4), 2299-2306. - ve42.co/Zhang2016
Micromouse Turn List, Keri’s Lab - ve42.co/MMTurns
Green Ye via KRplus - ve42.co/Greenye
Classic Micromouse, Excel 9a. Demonstrate fan suction, by TzongYong Khiew via KRplus - ve42.co/MMFanYT
Vacuum Micromouse by Eliot, HACKADAY - ve42.co/MMVacuum
▀▀▀
Special thanks to our Patreon supporters:
Emil Abu Milad, Tj Steyn, meg noah, Bernard McGee, KeyWestr, Amadeo Bee, TTST, Balkrishna Heroor, John H. Austin, Jr., john kiehl, Anton Ragin, Diffbot, Gnare, Dave Kircher, Burt Humburg, Blake Byers, Evgeny Skvortsov, Meekay, Bill Linder, Paul Peijzel, Josh Hibschman, Mac Malkawi, Juan Benet, Ubiquity Ventures, Richard Sundvall, Lee Redden, Stephen Wilcox, Marinus Kuivenhoven, Michael Krugman, Sam Lutfi.
▀▀▀
Written by Tom Lum and Emily Zhang
Edited by Trenton Oliver
Animated by Ivy Tello
Coordinated by Emily Zhang
Filmed by Yusaku Kanagawa, Emily Zhang, and Derek Muller
Additional video/photos supplied by Getty Images and Pond5
Music from Epidemic Sound
Thumbnail by Ren Hurley and Ignat Berbeci
References by Katie Barnshaw
Produced by Derek Muller, Petr Lebedev, and Emily Zhang
![[여고추리반3] 순한 맛 뒤에 마라 맛이 숨어있는 학교생활 | 4분 하이라이트 | TVING](http://i.ytimg.com/vi/6Wv9UEpFRNU/mqdefault.jpg)
![[Jeans' ZINE] 제2회 엉망잔칭 토론회 EP.1 데자부🤦♀️ | NewJeans](http://i.ytimg.com/vi/FN_KHxXZC0k/mqdefault.jpg)
Those turns are unreal, it looks like the mouse is simply teleporting across across certain parts of the maze
That one micro mouse surely do that once it hits 88mph.
@Dont_Read_My_User_Photo ok
Pardon the pun, but it is a-mazing how advanced these tiny robots have become, both in speed and intelligence. 😮
@Dont_Read_My_User_Photo TL;DR
That's what you get with 2 independently controlled wheels.
The section about the mouse choosing the longer but straighter path really struck home with me. Too often in map software, and even games with a GPS system, the "shorter" path will be taken, even though the longer path is actually faster when factoring in deceleration, waiting at stop signs, etc. It's really a fascinating area for optimization.
when it comes to people and autos in cities, the straightest path is usually better.
displacing a block for a better street can improve, the issue is top speeds are limited.
it is fun to optimize
oh yea thats good stuff i like to optimize while I am high on angel dust
Thats actually not true, modern gps factor in that too.
Totally agree. The 'trick' seems to be able to assign the correct speed/ acceleration times. Doing a 'flood fill' to find the route is one thing, but doing a second form of 'flood fill' where you label each position with an estimated time-to-goal sort of distance.
@@victormiranda9163 And we've all been there where we think, "I need to turn left at some point, which intersection would be the best place? At the light where there is a left-turn signal? Or maybe one block BEFORE the light where I won't be stopped by the light but maybe there will be a gap caused by the light where I can turn left even quicker?" Or do I just make three right turns? lol City driving, you quickly find yourself thinking about things like this and trying to remember what works.
Probably this might get lost in the sea of comments, but I just want to say that this video made me choose my first club at my university. We have an IEEE club, and it has a micromouse year-long project. I was so thrilled when I first heard about it. I am a CS major, but I've dabbled a little in electronics. I am exited about how it is going to go for me.
are you winning, son?
are you winning, son?
Lol it went better than I was expecting this first term, but I didn't win. I wasn't last either though. There are three more terms to go.
Although, I kind of dislike the software we're using. Bad and glitchy interface for the EDA part. The coding software lacks modern features such as multi line editing, static analysis, code suggestions and snippets, and has a dated UI.
For the EDA software I can only compare it to my experience using programming software since I don't know much about EDA software. Is it all so bad? The one we're using is Fusion 360. I'm honestly astonished it is paid software. Thankfully we're on a free student license though.
@@davidflores909sounds like you’ve found a passion project that could improve robotics and earn you a lot of money!
Good luck!!! Nerds FTW! 😁
Whoever does the animations to accompany the explanations for these videos deserves a raise! Those were top notch and absolutely vital to the effectiveness of this video!
Everything a computer/AI can do better is boring...
One thing I like about this engineering competition is that, since there aren't heavy financial incentives involved (like pretty much any other engineering project), people are given the chance to try whatever they want and be as innovative as they like.
Even with incentives, all they would need to be innovative is low costs for parts and work, and rules allowing multiple entries per participant (and no entry-fee)
Literally, if this competition doesn't show that capitalism doesn't produce innovation, i don't know what will
@@raymondqiu8202 You cannot argue that because A causes X, that B does not cause X as well...
@@raymondqiu8202 You don't understand the way capitalism does innovation. If there is big money riding on something, the capitalists will do quite well at figuring out how to get that money. It is only once they dominate a field, and risk becomes expensive, that they stop innovating.
@@jursamaj capitalism doesn't do innovation, people do. begging engineers to read one ounce of Marx
As an electronic engineer, this is one of the most epic electronic engineering vids I've seen. Thanks Veritasium
Same
Np i got u
So am I, I saw the length of this video and thought: "Eh not gonna sit and watch for 20 mins" but I got absolutely enthralled. Really considering making one!
IIT has a course called CS 102 you might enjoy
If _elevation height_ isn’t a “violation”… why not just launch a drone (aka: map~>process~>drive)💁♂️
14:16 I love the shocked reactions from the spectators.
Man, this is fascinating. Initially I thought this was a remote controlled competition, but after seeing how fast they were moving I knew that wasn't possible. It's really impressive what we can do with robotics these days, even on such a small scale!
Once you understand what goes into mouse navigation, this goes from appearing as odd nerd behavior to something genuinely impressive.
Yupp I think that's most things. That's why I love learning! Appreciate life!
well said. youre hired
and when they put machine guns on them and send them into tunnels after humans...
Same with any sufficiently advanced "odd nerd behaviour" tbh
Be that as it may, those are some shockingly weak mazes with **many** paths to success. Lets get some AI generated mazes in there where excessive diagonals are not included...
It’d be cool to see a maze with different elevations throughout.
or a 3d object which could have intersting shortcuts depending how the maze wraps around
That's exactly what I thought!!!
And opposite burms and different textures and bumpy sections
@@macallan3933 rally mouse
Non-euclidean mazes :D
The next innovation: Micro Mouse Tokyo Drift
Lol
Lol
They make you go faster, you get a boost when sparks come out
*Initial D soundtrack intensifies*
DURIFITOOO
Humans are absolutely beautiful. Both the people competing and over 12 million people on KRplus are invested in the idea of making a tiny robot solve a maze and it’s so random and came from just one person and now it’s huge. Sometimes I need things like this to remind me humans are pretty neat sometimes
Its just a shame that science and technology isn't the average human's main focus. The average human is more focused on materialism or climbing the social hierarchy ladder. If we were a mainly scientific species, the world would be such a better place.
@@DivinityAwakened In short, nerds should take charge of the world.
@@phunweng962 no, the nerd should not, or at least the top nerd should not. Sometime, more often that not, top nerd ignores moral for accomplishments.
Sure the mice are cool, but can we talk about the animations at 8:40? So impressive! No idea how they were made, but it really helped understand the concepts. Hats off to the team behind them.
Ikr its soo cool... Alsoo hello there blender guru you taught me blender thanks for that XD
Yoo
Ay the doughnut man
AGREED
Hello donut man
I'd love to see a 3D micromouse maze with all sorts of walls and ceilings and loops to drive on, using vacuum fans to stick to the surfaces
or use a drone micromouse
@@praveenmotamarri drons, but as small balls....without any fans outside, and can roll on wall to maximize speed when turning
I came here to say this. I see that it's already been said. So i second it.
Imagine all the possible Fosbury flops
@@praveenmotamarri there are drone obstacle courses...
It was neat watching the engineering evolve from just mapping the maze to taping off the wheels to increase traction
Thanks for this very interesting video.
Many years ago, Richard Browne, who worked as a technician for Bell Telephone, had seen an article published in the company newsletter that described and showed pictures of Claude Shannon's electronic mouse. Knowing that the mouse used telephone relays to control its motions and solve the problem and having access to scrapped telephone relays, he restored some relays and set out to duplicate the whole thing. The original published article did not detail how it was all done, so my friend figured it out for himself. I remember that the memory for each of the 25 cells of the maze area required two relays which recorded the direction the mouse had last left that cell.
Near the end of this machine's life ,somewhere around the late 1970s, I met and became friends with Richard. I saw the machine myself and was thrilled by how well it worked. Later, Richard went on to build marble machines, intricate wooden machines that allowed a marble to pass through various gravity-driven paths. Sadly, Richard passed away in 2013, but you can still see his videos about some of his marble machines. Although never completed, his grandest machine, called Marble Machine 3, was one of his creations described by Richard in videos here on KRplus.
As an electronics engineer this was one of my favorite projects that I have ever done. From the firmware, circuit design , algorithms and mechanical design every part of this robot is just pure absolute joy of engineering.
Do you mind giving me inputs? I am an EE student and I want to know what I need to know and any other things required to try make one
bro, are you bald?
as normal guy , i salute you electronics engineer. as a normal guy
I wonder how common extreme weight optimization is... Drilling holes into PCBs, using the lightest materials available, shaving them down to barely not break.
Surely that could get you a few ms due to faster turning speeds.
Also, how relevant are aerodynamics on this tiny scale?
Pin this
Man, seriously.. The guys behind the video editing and simulations in your videos are pure genius. Wish I could meet such guys to learn from.
🤔
They are Veritasium, mate
@@koenamh He's got a team behind him nowadays right? Did Derek make the animations? Does he still do his own editing?
@@Hugh.Manatee I'm 99% sure he doesn't
@@Hugh.Manatee Just look at the end of the description to see who made what in this video. Pretty detailed so i like it xD
In 1980 I was 15, bought my first electronics magazine. It had a micromouse article, that inspired me to get into computers. At 58, I'm a software developer for my own business.
This actually helps a lot if applied to food delivery, not even kidding. The shortest route is not always the fastest route. So I never even realized I was playing this game the entire last year, lol.
20:20 Imagine a giant cylindrical maze where the mice can go upside down. Or even a maze with loops like in Sonic, so the mice will have to account for more than a 2D map of the area.
I’m thinking 3D mazes in water or air mazes. But, surely, one can complicate it even more, adding rare shapes to it or even simulating 4 dimensions, building a tesseract.
It really has stopped being a maze solving competition, it's about movement execution.
Imagine quadcopter maze running.
mario kart 8
A Möbius-Maze?
As one of those who missed the podium of the All-Japan Competition this year, I can tell you that the level at which they are competing for the champion is on a completely different level.
one of them mentioned that he changed the optical rotary encoder disc from plastic to paper, making it 0.15g lighter!
what do they mean i cant attached nuclear reactors and rocket propulsion to my micromouse
Well on a micro scale that 0.15g could be something like 3 pounds relatively
So they have already swapped steel to carbon fiber screws or axles or straight to adhesives to save weight? Biggest issue i see with these advances seem to be that its money game to manufacture lightest parts witch require high quality tooling to produce well as having already wealth of knowledge on the robotics. Sure i can plan maybe even lighter mouse with things mentioned, but i'm decades behind on building the software to level where i could even compete.
@@Hellsong89 skill issue (couldn’t resist)
Is jumping allowed?
Great episode, thank you for doing all the research behind it. Love the explanations you gave, the algorithms behind it and the special moments over time :)
Gotta say my thanks as this video helped me create a first person maze game based on the maze circuit designs on the video for a university project and the panel/judge loved it with the addition of a horror theme through sounds, lighting, and objects in the distance. It was also interesting to learn about the micromouse competition while I was at it and I was able to replicate the diagonal movements the micromouse make on the mazes on my game.
Thank you again.
When the guy you were interviewing said "you come along one day, you see everything and you go "huh. that doesn't look to hard, i could do that,' but then you find yourself sucked into a deep and rewarding hobby" i felt that in my bones. I got into rc planes much the same way. Honestly watching this micromouse thing made me think the same thing like huh i bet i could make a mouse..
you are right, i felt his words too! i have a similar story; ive watched a lot of rc planes but it looked so damn hard and the cost involved is toooo high so i never ventured into it. however when I had the chance to attend a uni technofest where I saw contestants competing with their soccer cars, it clicked - this doesnt seem too hard! And thats how i got into making my first rc soccer car using arduino uno. now this mouse vid makes me think, how to even start coding such a thing!?
This is how I got into slot car racing. Just go along, enjoy the weekly racing... go to a national race competition, finish in the bottom 10%, get the bug more... a few years later, routinely entering 6 and 12 hour races with an annual 24 hour (as a team, not just myself o_o)
Still want to get into RC aircraft, but £££ of course. And I want to try the fan concept from micromouse in a slot car first, to see if it even works with our tracks (wooden tracks with a routed slot are often smooth, but the majority of my races are on plastic track that has imperfections... the smooth plastic sheet from the downforce demo in this video would be a dream to race on!)
Yeah but like rc planes, you really have to derive your own joy from it. Maybe it's different where you live, but people are kinda a-holes about the rc aircraft hobby. They either don't want it around, or they've forgotten more than you'll ever know. Maybe it's like that with everything.
I thought to myself "That is the driving force of humanity lol" when the guy said that.
Timestamp?
It's amazing how such a simple concept as a robotical mouse running in a maze can have so many implications and thought put into it.
Maybe it is simple that enables competitors to come up with original ideas.
Cannot believe i watched an entire documentary about small little robots having the same g force as formula 1 cars and going so fast you can barely see it, Loved it.
I havent seen this pop up here yet but Theseus is a pretty good name for essentially the precursor to AI.
The Theseus' Ship paradox goes like "if we replace every single piece of a ship over time, can we still call it the same ship?"
Because if you compare the final version to the first version, literally everything has been swapped out and yet, on an individual basis, the majority of the ship has always stayed the same. We build out home tower PCs on the same grounds.
And while the core idea of a mouse solving a labyrinth has remained the same, pretty much every component of the original Theseus has been improved on, swapped out or both.
am i the only one who had a big wide smile throughout the whole video? i am just amazed and fascinated by simplicity of task but the ferocious ingenuity of the competitors. just loved it
same
lol same
Same
same
I kinda wish he'd segued into robot-sumo. The robots and many of the strategies are similar, but they need to tackle a dynamic problem (push the other robot out of the ring) rather than a static one (navigate an unchanging maze).
My respect to all the previous engineers of the past whose mouse was really slow, their consistency to push this competition further paved the way for today's modern engineers. This is one of the prime examples of what humanity can achieve while working together generation by generation.
@@mahyarshokraeian It's part of my humanity to claw my way out of eating nothing but scraps, and onwards into engineering.
Saying that feats of science is a waste invalidates people who actually escaped the poverty line through science. It's like telling me that I need to go back down there just because I'm able to feed me and my family now.
Instead of being derogatory to science, why don't you just inspire people to aim for this knowledge so they too escape like I did.
People inspired me to climb, and so can you.
Oh yes, and by playing playfully !!!! While doing the most serious of all engineering works.
@@mahyarshokraeianyou don't have to choose. You can do both at the same time.
@@mahyarshokraeian I'm sure you spend every second of your life dedicated to making the world a better place and never waste time on any kind of entertainment and you only spend enough money to survive while donating the rest to charity.
@@mahyarshokraeian what do you do to stop world hunger?
I love the idea of this. It's a really cool way to involve science, mathmatics, robotics, and fun into a really cool competition.
16:13 Geometry Dash is a 2013 game made by popular indie developer Robert Nicholas Christian Topala, also known as RobTop.
the way they manage to maintain a perfect distance from the walls, and go SO FAST is insane
Not really. Keep in mind that the microcontrollers in those robots can do hundreds of millions of computations per second. If any uses an FPGA instead, that could be an even bigger number. From the robots' perspective, it must "feel" like driving at 0.01 mile per hour.
@@TheNefastor it’s just incredible to watch something so small maintain such precise control from the perspective of someone who has very little experience in robotics.
@@TheNefastor using FPGA's was also my first thought in reducing the computation times dramatically. I don't think that any of the winning mices are using a microcontroller.
@@Craftlngo I wouldn't know, but don't underestimate MCU's. The fastest STM32 runs at 550 MHz last I checked, that's plenty enough to run this kind of challenge.
@@TheNefastor How about the momentum, power cut-off, and wheel friction, do they calculate all of them as well? I wonder if let's say the sensor and computer can decide the next step fast enough, will the output of the movement be straightforward?
That British accent "OH YOU SNEAKY LITTLE-" when it turned diagonally for the first time is hilarious to me
"Oh you sneaky devil you, you cheeky bugger" 14:18
i love the peoples reaction to the first ever micromouse to cut corners, you can hear everyones amazement at that first turn.
During my Engg college days, we guys did Pick & place bots for inter-college competitions. Those bots had to navigate similar obstacles, pick the object like a small TT ball, or a cube, then place it somewhere else. All controlled remotely. Wired or wireless, anything was fine. And a couple of years later, Bot Hockey too. These bots had to play Hockey.
These kinda videos are reminiscing my college days, where we guys used to spend nights together in creating these Bots & playing with them. It’s soo satisfying & brings smile to my face. 😀 Thank u ❤
Pretty impressive. Going from "this is boring" to "I bet I will be the one who come up with the next big innovation on this great sport" in less than 25 minutes.
Best of luck on your journey. ⛵
Check _superiority complex_
you're a candidate to go from "this is bs" to "I'm the best example ever" in 11minutes
Calling that mouse “Red Comet” is such a great touch. It’s a reference to the nickname to a mobile suit pilot who was 3x as fast as the others in Gundam!
Spotted that too, I love that reference! 😁
If they have a Char Aznable mouse in the competition, Mighty mouse = Big Zam.
red always go faster
oh its not about that guys dog then?
Along similar lines, the original competition LeMouse 5000 refers to the french 24h endurance car race at Le Mans (as the french use "souris" not "mouse" for computer & other mice).
This is so interesting!! I didn’t even know they had such competition but glad I got to know this exists!!
The story telling on this video feels great! I have been following this channel for many years now. But how captivating the videos are still keeps increasing. I just watched 25 minutes about little mouse-like maze-solving robots, but it felt like 5 minutes.
As someone who has participated in robotics competitions, it's so emotional, it's like horse racing, but you have spent months building the horse with everything you have. Just being there with all these people, seeing their genius solutions to the problem, it's so much fun. I would truly recommend it to anyone
which league of competition? I was on an FRC team in highschool!
@@stuchris I participated in student robotics, thats a competition in Greate Britain, but my group also went to EuroBot several times
How much technical knowledge do you need to get started, can you do it without an engineering degree?
@@yune1000 you can literally do it in middle school with no prior knowledge
@@yune1000 absolutely, I was in highschool, so I had no degree at all. I think you can learn most of what you need on KRplus. Soldering would be a good skill and some basic knowledge of how coding works. Everything else will come with time and practice
even including fans for suction? these guys are insane, the amount of work put onto this
When I first saw the footage I wonder where they were getting that much traction from and my assumption was magnets (and the maze was built on a metal plate) but doing it with a fan is a far cooler solution to that problem. Self contained as well.
@@truepennytv i thought it was similar to the road used in drag races, the road is pretty sticky, but yeah you're right
And what is so cool is that as long as you stay in the general rules that keeps the spirit intact, no one will tell you "no you can't use that it's too good"
F1 teams were experimenting with that technology in the 70s (as far as I remember), but such fans were prohibited by the technical rules after only one season as fan failures in turns or cars hopping over curbs was devastating and even fatal.
When you think about it, it is actually quite obvious to every car geek.
There have been tons of ground effect cars, including the chapparal 2j (with a fan), or in formula 1 the lotus 78/79 (with skirts), or the Brabham (with a fan, but different), etc.
So IMHO, yes, surprising, but also obvious at the same time 😂
Impressive and scary how quick and accurate they can move.
never herd of such a race, but was rather interesting to learn about it. Thank you for this little nugget of information.
This video was a roller coaster: First stunned because I thought they have to solve the maze in their first run, then disappointed because they didn't, then amazed again about the speed of these things and the engineering done to achieve that
I love that Red Comet got there faster than the other mouse by actually doing what racing drivers do and taking the *racing line* to the goal, ie. maximising top speed through the straights and minimising speed loss by taking fewer turns. Brilliant.
The fact that that consideration doesn't seem to have occurred to any of the other competitors before then tells you something quite deep about knowledge: you don't know what you don't know. There will always be 'unknown unknowns'. Saying 'we've reached our limit, we've solved all the problems there are in this subject' is a failure of imagination.
i really thought "unknown unknowns" wasn't the right word to use but u prove it right (or wrong?)
How do racing drivers take fewer turns if they're racing the same track?
@@FancyUnicorn Racing "line" is important because the track width is usually 2x - 5x the width of the car. The most efficient use of available grip comes from maximizing the radius of each turn (although there are exceptions) by moving across the track.
For instance, the "racing line" entering a left 90-degree corner would have the driver start on the right side of the track approaching the corner, move to the left during the turn while striking the inside left apex, and end up back on the right side of the track on corner exit. This "line" maximizes the turning radius, and is much faster than staying on the left side of the track for the entire turn.
@@nativenugget major boondocks vibes
@FancyUnicorn they don't take fewer turns. They take more efficient ones.
Wow that was a very interesting, unique and whimsical documentary. Thank you. I left this feeling happy
MAN!!! I have never been more interested in a video. I competed in a Lego Mindstorm maze completion back in 2003, when I was in highschool. The point you bring up about dust was a factor, BUT the biggest factor for us wasn't dust... it was the charge on the AA batteries... You never mentioned that, but you definitely should of. As well, we won the competition, entered the robot in the science fair and wone an award from Intel.
Woah! When I saw this video was posted, I walked straight out of my office 30ft to tell Dave Otten (from the video) that it was posted. He was quite excited!
Did you make diagonals in the hallway?
@@ChemEDan between lab benches, yes. Gotta be efficient
seriously under rated comment!
Ah... 30 feet isn't far enough to justify using a portal gun, I get ya
I don't think people understand how big this event is in Japan.
I was teaching robotics in Japan and a student said that I should come to this club and check it out. It was the Micromouse club I I was shocked at how awesome their hardware and software was.
I had to take a serious look at my viewpoints on robotics after seeing this subculture.
But do you have to be a virgin or can anyone compete?
but why do girls somehow get weirded out when I wink and tell them I can show them my micromouse in my bedroom? are they not into robotics as much as I thought?
@@laimejannister5627 they prefer at least an average size mouse
@@taoarg9000 one time I said I could ask my bros to bring some big ones over and we could try them together then she just left. haven't seen her since.
@@laimejannister5627 Did you try telling them that you can finish in under 10 seconds?
This is such a testament to engineering. I'm always impressed how generationally, or year to year, engineering evolves in complexity to become more and more precise and innovative far exceeding what the original concepts could have ever predicted.
This was absolutely fascinating. I never knew this existed. Thanks.
I'm so grateful that I live in a time where I can get this level of information from my couch for free. What a time to be alive.
Facepalm
@@dr.angerous why?
@@Ak-us3sh Probably some ancient alien believer.
Every time I see that phrase I think of 2 minute papers
@@Tp.123- The planet is dying but what can one man do about it? The only thing that can be done is taking down the tyrants at the top.
The really interesting part of this for me, as a motorsports fan, is the fact that so much of the innovation has followed a similar path to motorsports.
The famous Brabham BT46 "fan car" was a F1 car that had a fan attached which sucked it down to the ground, providing greater downforce and cornering. It debuted in 1978 and was banned after a single race (which it won).
I was thinking about this exact thing. I like how micro mouse doesn’t change the rules when a certain change becomes commonplace or too dominant unlike in Motorsport. There’s always room for innovation.
@@pobrecitossb7450to be fair, innovating on a huge vehicle is far more expensive than innovating on a mouse sized robot.
No, no, that was just a cooling fan. 😉😄
Thank you. I was trying to remember where I'd heard about the car with a fan for suction to hold it to the ground.
@@tthurlow thats not really the issue that keeps motor sports from doing crazy stuff like they will try in this competition. While yes innovation in motor sports is expensive, we've seen over the decades the insane amount of money companies will spend to innovate to win a race.
The issue with motor sports is the fleshy meat bag operating the vehicle. If a little mighty mouse robot runs into a wall at 50 mph, catches fire and goes pop, it's kind of funny and everyone might be out some money. If an f-1 car goes airborne doing 250 mph into the bleachers because Mercedes was allowed to do whatever, you could easily see a large amount of human casualties.
A great example is the old group B rally circuit where there where minimal rules for the manufacturers. It got wild, the cars where insane, tons of innovation and lots of death and injury for the short time it existed.
this is way more fascinating than I thought it was gonna be.
Unbelievable that this is a Veritasium video. No condescending attitude, no gimmick, no clickbait. Wow. Awesome Ve, love this format, please release more like this!
Sorry to say this but it’s true lol this is better than when he’s on camera
I would love to see a layered maze with multiple floors and ramps, like a parking hall. Bottom is the start and top is the goal
And they need to make the problem more complicated again - like adding in the free-standing walls. How about some curved walls, or pegboard holes in the floor, or rough surfaces, or transparent walls, or curtained-off short-cuts?
Or add "cats" that block parts of the maze as they follow predefined paths. So the mouse has to probe the movement of the cats as well as the walls to find the best path.
Doors/gates
Opening and closing gates like in fall guys
Add a lava moat filled with fire alligators
If you combine maze-solving with "battle-bots", the introduction of multiple bots into the maze trying both to reach the end first, and to destroy their opponents would be very interesting!
plot twist, a maze-creating robot entered the chat
@@danosdotnl now we can introduce our hero "maze runner"
the world needs this
Like king of the hill for mice
PAC MAN style , bring it back
I wonder if using small electro magnets to shift weight to different sides could help knock/slide/push the mouse internally to increase turning control making sharper faster turns and even help with recovery from those turns.
I'd like to see a hexagonal maze or better yet, a maze with curvy walls in these maze solving competitions.
Some History: The first Micro Mouse competition was won by former colleagues of mine at Battelle NW in Washington State. They had a 3-step algorithm - 1) Random Walk through the maze, 2) Explore every square (and walls) not encountered during the first run, 3) Compute and drive the shortest (distance) path. Of course the early mazes were much smaller, and they were also less complex so that a mouse with no smarts could execute the wall following algorithm. The fastest (time) mouse the first year used that technique. The microprocessor they built was from what Intel called "Floor Sweepings" - fully functional, but cosmetically defective chips. These incredibly talented engineers were the first in our department to put together an embedded system as I remember. It was a FANTASTIC place to work full of innovative and amazingly personable people. Best job I ever had! Our group's claim to fame 5 years later was to create the world's first self-contained Rubik's Cube solving robot. Just like the first Micro Mouse, Cubot's time of 2:40 has been eclipsed many times over. Still...there's a nostalgic feeling that grabs me every time I see one of these mouse competitions...
woah. that's great
That is actually a cool story. You must be really old though
During those times, what did you guys envision future technology to be like?
@@besterspieler2285 If you're lucky, you will be too one day. Respect your elders.
@@forthebirds4 what he probably meant was that since he has witnessed all that, he might have been a young engineer then, so definitely he may be old, there is nothing disrespecting in asking if someone is old or not, did he de mean him ? What's your issue, what's hurting you so much, calm down bro, with that attitude you are going nowhere....
It'd be interesting to add some curved sections to the maze and see how that affects the routing algorithms. It looks like the mice can already handle them mechanically.
CURVES SECTIONS?!?
Something tells me that could either make or break the algorithms.
Think about the flash fill method... they use a grid to map the maze. Now how would that grid work with a curved section?!?
@@00linered I guess you could either subdivide the grid further, or maybe work with floats?
@@00linered wouldn't that be the new challenge?
Or bridges
@@jayathranps1319Exactly, transform the sport! Curved sections would be so much fun to watch.
This is great! It seems like these micromice are helping with future designs in robotics, which most likely will have much more important applications than just solving mazes!
Those downforce results are amazing. It makes me wonder if they should include an inverted maze category. Imagine a maze with a ceiling and no floor.
Or a 3D maze, a big cube where you have to get to the center
Ngl, an integration of a multi-level and inverted maze would make this so much more complex, yet exhilarating to watch
@@teslatrooper ooh put a nice little cmos and gimme a vr :D
@@foxgaming76yt24 Drones will get there
But why there are no curves and roundabouts in the maze? Weird and sad.
Wow, I'm almost 40 and I remember competing in this kind of competition back when I was 15 and studying for my IT GCSE. I remember spending hours tweaking the motor commands to the main 2 wheels to be able to take corners as quickly as possible and very crude attempts at a maze searching algorithm - I think I made some terrible combination of trial and error and "always keep your hand on the left wall". One of my best memories of that class :)
wow
Amazing. Never heard of this before seeing this video.
@@MonikaDudek-sw8piPlease am new to this and I've incurred so much loss in investing.
Am so happy my financial life has changed ever since I knew Mr Alan Hernandez I've been earning over $20,600 every week.
I'm also happy to start trading with Mr. Alan Hernandez, who has time to monitor your trading account with an Expert is the best strategy for novice and busy investors.
20:45 that’s WILD!
Crazy crazy crazy
This is fascinating, thanks a lot for the video 🙏.
I love that they are allowed to experiment and add new hardware to the mouse; it will be great to check back in on these competitions in a few years time and see what innovations they have come up with
I think this is crucial to the event's longevity. "Solved" competitions are only solved when there is no more room to iterate within the ruleset, so flexible rules are the best way to foster innovation.
Maybe in the future there will be variations of the game involving 3d mazes and other types of obstacle courses
It will be embedded with the brain of a 13 year old high school sprinter with angst issues, forced to run the race for eternity to save humanity from the rat race.
As someone who led a micro mouse team back in undergrad, this video is extremely well done and interesting. Thankful this video exists.
i’ve decided you didn’t and are wrong
@@TheFakeDingieWingie bro is disgrace to hampter lovers
@@memehamsterr one of these days your shoes a will tie themselves.
The sections / structure / narrative of this video, in order to be exhaustive about the evolution and multiple aspects of that sport. Once again, second to none. Great content, sponsors very much on point, and at the end of the video. Amazing channel.
A true Veritasium classic!
I love it when I stumble upon this part of KRplus where I didn't know such a thing was even a thing and glad I discovered this interesting subculture.
I love how the video is building up the tension of the Japan competition of Utsunomiya trying to beat first place.
It's just such a treat to watch
also Red Comet being named after a famous anime character known for being unusually fast.
Got me subconsciously rooting for that mouse to win.
@@nicoliedolpot7213 Sieg Zeon!
A submarine version would be really interesting and get even more fluid dynamics involved in the problem.
Submarine would also be 3D maize rather than 2D
They already do this. Its called RoboSub competitions
great!
@@asandax6 What if there was a 3D version that required flying? That'd be a challenge.
@@Doctor_Yuri but that’s not a maze competition
This inspires me. There is no need for any fancy reviews. Simply put, this inspired me.
The explanation on how that works together with the history of it is FANTASTIC!
That just made my day! 😃👍
This videos was absolutely stunning. You took a sport nobody knew about and turned it into an amazing video. I also wouldn’t of understood a thing without those visuals. Hats off to the team.
My head was bashed in as a baby in the NHS hospital I was born in leavingme withlearned disabilities with government involvement
I'm sad that this is the first I'm hearing of this amazing competition. Thank you for bringing this to our attention!
It loses novelty fast. I've watched before
*His method surprises me. A Friend that I referred to him, just received €50,150 profit after 7days of investing.....I became jealous,...Lol*
the flood strategy is like the when we follow the compass in obsidian games ( skyrim etc ): trying to go straight the objective, getting around obstacles on our way
This reminds me of an old book in Reid Kerr Engineering College I found 1970 +/- one year. Logic behind mechanical mice.
It described 3-wheel self propelled mice which would not fall off a table.
There were two tables considered. One had a slightly raised edge, the other had none. The latter had two sensors, one to detect a drop ahead of it and the other to detect a collision.
A collision in either or a detected drop in the latter type would trigger a right turn.
Now that I have seen computer mice I wonder if Xerox Labs had any input. But the book looked quite old, although 10 years would be old.
Absolutely fascinating. It's a real shame this stuff isn't aired on major outlets.
They should have an ESPN alternative channel, with all these more special competitions. This, wife carrying, speed lumber jacking, eating competitions, pumpkin throwing, dodgeball.
@@Binaryrunt So a permanent ESPN Ocho?
KRplus is a major outlet
@@Binaryrunt can you include tag in there or is it already being broadcasted well enough
So many things to air literally
Im a motorsport enthusiast and got really excited when they announced Roborace. Sadly it never caught on and became a joke before disappearing.
Seeing those robot mices racing in those mazes really is awesome and I'd happily watch an AI race with full size cars and tracks.
I'd be able to enjoy crashes without guilt!
@@DizzyDisco93 yeah and since there wouldn't be no human inside the vehicle, safety would be limited to spectators and we could have super light and fast cars !
In all actuality you'd watch it once and probably never again because it would be incredibly boring. At top tier racing, once all the cars are the same from hitting the limit of technology and/or budget, the only real dynamic factor is the human component.
@@zwan1886 there's always a human component, and already today the bigger factor in most racing events today is more the human engineer rather than the human driver imho.
@@ThomasKafer Then you don't know much about racing because they put caps on what is allowed, and when there aren't the races are won by whichever organization has more dollars, so it's just pay to win which doesn't make for an exciting spectator sport either.
My favorite cameo for Micromouse came from an early episode of the long-forgotten but groundbreaking sci-fi TV show, Max Headrom. Micromice competitions were a part of the curriculum of a school for gifted youngsters that were involved in the plot of an episode.
Is that the thing Seth Meyers hates?
I didn't even knew this existed. Thank you for this knowledge.
I did the micromouse challenge 20y ago on robotics class at university. This was a good trip to memory lane but also amazing to see the current level of all participants! Truly outstanding!
It's a little mind blowing. I still see optimization I is there because the mice are not taking advantage of racing lines as much as they could. It is Incredible what these little robots can do.
@@tmi1234567 wdym by racing lines?
@@Martineski bit like racing colours (racing red) but for lines
Looks like the Japanese are dominating this field/challenge.
@Martineski Racing lines are the theoretical line around the corner that maintains the highest average speed and least amount of time. This is where you hear things like, hitting the apex, where you go from out wide into the inside edge to take the widest turn you can without traveling excess distance.
These mice seem to be taking mostly straight lines equidistant from the walls to avoid crashing into them, diagonals excepted. To optimize, the mice would need to hug an outside wall, than turn in a bit early to just kiss the inside wall, then barely miss the next outside wall.
I was in the 1990 UNSW Micro-mouse team in Australia. The previous team had just transitioned from stepper motors to DC motors. The high speed DC motors caused frequent wheel slippage making positional calculations difficult. There were a lot of challenging problems to solve but it was really fun. It's good to see so much progress since then.
just happen to suggest in my feed and this is one of the awesome video suggestion that came up. amazing
this competition that i have never heard about is so charming i love it
Ive never once studied robotics but it seems to me that this sort of thing would be a great introductory course to the subject
Sorry bro, but this kind of thing is so far away from introductory, introductory robotics are like: open and closing a gate, or lifting up some wheight with a motor. Actually doing robots its on the midterm of robotics, and competitions like this are endgame things.( Sorry for the possible typing errors)
@@klenom112 I think he meant it in the way of being introduced to the potential of robotics during introductory courses. Show students the possibilities, let them imagine the what ifs, and then the basics begin.
@@anotherdayanotheranimation Excelent point man, didn't saw that way, it is a actually awesome way to introduce robotics.
Its actually really not that far out there to use this as introductory robotics, even in a practical sense! Before I went off to college, I participated in a highschool robotics competition that McGill University hosts, which does exactly this. All the equipment is standardized with a few customization options, and over the course of a couple days they introduce the different features and how to write code on the arduinos that drive the little mice, and then you have a sandbox day to try different things with different mazes before you submit your final version for the contest. They use very simple mazes, and you have very few options for sensors, but in my opinion it was the perfect level of challenge!
In my university it's a class they teach you to code and then at the end you build a micro mouse or a robot that does something like following a path then lifting a can or moving some servos to carry stuff. This class is on the camputer/electrical engineering idk if there is something similar on computer science since they only do coding and stuff while we do hardware and some coding.
I did this as part of my electronics university course and it was a lot of fun! Didn't end up with anything groundbreaking, but it's great to go through every stage of design and prototyping and create something that actually solves a practical problem, even if it's a small one.
Gathering Information, processing decission, controlling motors -> milliseconds
Opening Email Client -> 10s
Love this video and topic. I enjoy watching this video every few years.
It's amazing to think about how intelligent systems can approach a structure like a maze. While watching this, I started thinking about how I approach maze like structures in games like Skyrim or Doom. I think most gamers probably have a strong intuition for maze navigation, but have no idea of what that strategy is in logical terms.
To sum it up:
If you go to the finish line before you know the ENTIRE maze, you are disqualified
It feels like there’s only 1 equation and about 4-5 “if…then” procedures at most. The rest is controlled by a gyroscope automatically.
Best maze tip almost everyone who solves mazes knows, stick to the right wall unless you loop yourself. Some mazes do that and I hate them. It's very nasty how some of them are so complicated and always just loop after you explore such a deep branch that you have to return because it was one way.
It's human intuition, machines are and possibly will forever have problems with.
No AI here. These are hard coded solutions. Those solutions evolved on wet computers.
Even including fans for suction? these guys are insane, the amount of work put onto this🤯
As seen on some crazy racing cars from the 1970s
excellent "documentary", much appreciated. Your stuff is always high quality.
19:50 so in Formula 1 that is called "ground effects" lol. Robot mice using ground effects to go faster. All we need is a Mice GP, and we are all set.
I just spent 25 minutes engrossed in a video about tiny robots trying to solve something you find in the Sunday paper. This channel continues to amaze.
You forgot to add all the times you rewound so you could see it again...
I see what you did there. A-maze
A-Mazing!
@@hisober 100% unintended, but sure, I'll take the credit 😬
I get that they try to go fast as possible, but going right always solves a maze. Try it yourself. Don't look ahead and always follow the right wall.
I love these kinds of things. It feels like you met someone 10 years ago and thought their hobby was kinda interesting. Then you come back and they have taken it 1000x further than you could even conceive of 😂😂😂
wow, this is such a great video! Loved learning all about the robotic mouses and the competition.
Oval circuits are not like other race tracks. Some tracks don't necessarily have "banked" turns, in racing it is referred to as a corner having camber, as the wheels of a race car utilize negative camber angles to optimize the contact patch when cornering, negative camber is when the top of the wheel is closer to the center-line of the car than the bottom, improving lateral traction when the suspension and wheel is loaded during cornering.
16:14 geometry dash reference
True
**Slaughterhouse intensifies**
@@mikoGD_**FEW WILL STOP TO HEEAAARRR**
hi
death corridor