Triangle Tank Version 2

Подписаться 1,3 млн

Просмотров 309 тыс.

50% 1

Go to brilliant.org/... to get a 30 day free trial + the first 200 people will get 20% off their annual subscription.
This is version 2 of my omni-directional tank.
The tank is omni-directional. That means it can move forwards, backwards, and left and right, or diagonally, or any combination out of 360 degrees. It can also turn on the spot.
Each track has little wheels mounted in each segment which allows them to slide sideways. This is really similar to other omni-directional robots I’ve built which use various types of omni directional wheels.
This seemed like it worked well, but it was really bad at climbing over things, which is what tanks are generally supposed to be good at. Two tracks at an angle like that just slipped on the edge of most objects and it couldn’t get traction. Also, as the front of the tank was lifted up, the back edge was pivoted off the ground, which didn’t have any drive in that direction anyway.
The main suggestions to improve this from the comments on the original video were to make the track style a trapezoidal shape like a traditional tank, and also to have each section able to pivot so each track can comply with the terrain.
Also at least one person suggested moving the little wheels in each track segment to the outside of the tracks so that they can help grip better on the corners as the track approaches objects.
CAD & code: github.com/XRo...
You can support me on Patreon, join my Discord, or buy my Merchandise:
***************************
Discord: / discord
Patreon: / xrobots
Merchandise: teespring.com/...
***************************
Other socials:
***************************
Instagram: / xrobotsuk
Twitter: / xrobotsuk
Facebook: / xrobotsuk
***************************
Affiliate links - I will get some money of you use them to sign up or buy something:
***************************
10% off at www.3dfuel.com/ - use code XROBOTS at the cart screen.
Music for your RU-vid videos: share.epidemics...
***************************
CAD and Code for my projects: github.com/XRo...
Huge thanks to my Patrons, without whom my standard of living would drastically decline. Like, inside out-Farm Foods bag decline. Plus a very special shoutout to Lulzbot, Inc who keep me in LulzBot 3D printers and support me via Patreon.
HARDWARE/SOFTWARE
Below you can also find a lot of the typical tools, equipment and supplies used in my projects:
Filament from: www.3dfuel.com/
Lulzbot 3D Printers: bit.ly/2Sj6nil
Bearings from: simplybearings...
Lincoln Electric Welder: bit.ly/2Rqhqos
CNC Router: bit.ly/2QdsNjt
Ryobi Tools: bit.ly/2RhArcD
Axminster Micro Lathe: bit.ly/2Sj6eeN
3D Printer Filament: bit.ly/2PdcdUu
Soldering Iron: bit.ly/2DrNWDR
Vectric CNC Software: bit.ly/2zxpZqv

Опубликовано:

26 сен 2024

Ссылка:

Скачать:

Готовим ссылку...

Добавить в:

Мой плейлист

Посмотреть позже

Комментарии : 444

@jamesbruton Год назад

Join my DISCORD! discord.gg/fc6MedG7eW INSTAGRAM: instagram.com/xrobotsuk TWITTER: twitter.com/xrobotsuk MASTODON: fosstodon.org/@XRobotsUK

@justpaulo Год назад

8:00 I don't understand your math. I believe you are wrong. I think the horizontal track must rotate 2x faster than the V shaped ones (and not 1.5x). PS: I keep trying to paste a link to a picture showing my geometry argument, but the comment is automatically deleted.

@jamesbruton Год назад

@@justpaulo Well it works and that's the way I've built all omni robots

@justpaulo Год назад

@@jamesbruton I know, I see all your videos... You do an impressive work by the way! On those other projects I also commented something, or tryed to, but b/c of links the comment gets deleted. I am going to try to share with you a picture of my google drive that I think it shows my point. Basically lets imagine that @ 7:42 the robot is moving right with speed 2. The right track points in a 60º angle and the little wheels in that track point in the -30º direction. For that track to move smoothly 2 to the right, the little wheels need to move √3 in the -30º direction and the track to move 1 in the 60º direction (it's the famous 1, 2 , √3 triangle). Therefore the speed of that track is 1/2 of the horizontal upper one.

@jamesbruton Год назад

@@justpaulo No it isn't. The horizontal track runs faster otherwise it spins as it drives.

@justpaulo Год назад

@@jamesbruton But that is what I'm saying.... I'm saying the horizontal track must go 2x faster though (not 1.5x).

@timixretroplays Год назад

An idea for more intuitive omni-robot control: Put a compass in your controller, so it knows which direction you're facing (which it can assume is also facing the robot). Put another compass in the robot, so it knows which way *it's* facing as well. Say you're facing north, and push the stick to the right - the robot will figure out it has to travel east, and work out what power it needs to put to which treads to go that way. It'd be more intuitive because all you're doing is telling it to go left, right, away from or towards you; you could also add automatic rotation, so it can always wind up facing whichever direction gives it the most efficient/powerful travel (perhaps with manual rotation so you can still wiggle it left and right to get over extrusions). It's an omni-directional robot - you shouldn't have to care which arbitrary direction is "forward", that's its job! :)

@nerys71 Год назад

Oh yes some of the lower end drones have a similar feature called heading lock it doesn't assume you're facing the drone but it always assumes you're facing the same direction which was surprisingly effective once you figured it out

@edmills9736 Год назад

@@nerys71 "Headless" mode on drones.

@jackmino729 Год назад

You could probably combine this with some other automation so it can follow you around

@TalynStormcrow Год назад

Great minds think alike, I just suggested similar without seeing your comment first. 😊

@Beef4Dinner22 Год назад

For anyone wanting to try something like this, the sensor you would be looking for is a magnetometer (specifically you would probably want a 3 axis magnetometer), which measures magnetic fields and can tell you which way is magnetic North.

@alan_felt Год назад

Do you have a system that manages all your prints and printers? Seems like a lot to keep ...track... of.

@its_nihar Год назад

Yes I have had the same doubt. Yeah basically if you do em all manually i think they'll probably take a month or you don't sleep 🤣.

@A_Bl4nk Год назад

THATS A GOOD ONE

@Spiker985Studios Год назад

I'd presume Octoprint would be able to do so, but alas I haven't used it

@Esablaka Год назад

Not sure about lulzbot, but several of the large 3d printer manufacturers offer "Printfarm software/support". There you basically have all your printers piped into a program and whenever there is an object to be printed the printer who currently is free will start automatically (and which has the configuration like filament etc.) . There are even manufacturers with hardware options that clear the bed automatically and store the parts in a bin so you can print continuously without needing to interact with it at all. Creality, prusa, flashforge etc. all offer this functionality as far as I know. One can also set it up oneself via octoprint or a software like Simplyprint .

@HarveyShankar Год назад

I've heard he keeps them in a tank.

@RC_Engineering Год назад

Amazing job taking all those suggestions into the new design. I guess what we learned was, a 3 tread tank has more issues than benefits

@ulysses4536 Год назад

I would argue that we needed 2 less iterations out of total 2 to prove that it's not a good tank design.

@anonymouspersonthefake Год назад

@@ulysses4536 it still might make a nice remote control toy

@ronnetgrazer362 Год назад

@@ulysses4536 Waste of plastic, time and network capacity. I'm done with this channel.

@DashieDasher Год назад

Your videos are a wonderful example of how 3D printing can be used as a prototyping tool

@club2772 Год назад

Really enjoy how you casually critique your work. Similar to skepticism in a good scientist. It’s cool to see your reworks of projects, that process alone is fascinating to see. Keep up the awesome videos! 👍

@CheeseJuggernaut Год назад

Holy, just started getting recommended your content again. Can't believe it's been 7 years! Most consistently great content!

@BlackSkull1984 Год назад

Bro, your robotic inventions will make the world change

@nnehila Год назад

I think those custom treads with just 2 tracks total would work incredible. One of the main benifits of tank tracks is that it can turn on the spot (ie like omni wheels). The third track seems incredibly silly after this video haha. Now a 4 track version (2 front, 2 back) would be what would be awesome for James to scale up to full size

@Mysda_ Год назад

Then it wouldn't have any sideway control with just 2 parallel tracks.

@stalkerfromvoronezh4493 Год назад

@@Mysda_ But 3 tracks fuck up all idea of tank tracks.

@Mysda_ Год назад

@@stalkerfromvoronezh4493 I agree, feel like 2 tracks with mecanum 45° wheels could be a better option. But I don't know if that can turn or not

@2009dudeman Год назад

@@Mysda_ That wouldn't work. mecanum wheels require groups. Two tracks would behave very strangely. It could drive forwards or back and pivot in one spot, but any attempt at differential drive would result in unwanted side shift in addition to the turn you intended. We did try a 3 wheel mecanum setup and it kind of worked, but it was not very good and had some strange behavioral quirks that made it essentially unusable. You can have more than 4 independent wheels, it just makes the math more complicated and you can no longer just use the single drive equations for each motor (sin functions) you now need to keep track of two coordinate systems and calculate the vectors for each wheel individually to determine the required wheel speed.

@Mysda_ Год назад

@@2009dudeman Any place to see that? I'm interested! I was really wondering because screw-propelled tanks seams similar enough to the concept, and they actually are able to go forward/backward, rotate and also strafe with only 2 motors. Also yes, it can get pretty complicated. I was brainstorming about a neural network capable to adapt to any omniwheel setup on the fly a while ago, never ended up making it.

@nerys71 Год назад

If you want to maintain that sort of a profile what you need to be able to do is to tilt the tracks but not the way you're tilting them You're tilting them on their rotational access coming directly out of the body instead you need to tilt them on pitch so when you are driving using two tracks you need to pitch the dead track so that it's angled up in the direction you're traveling and when you reverse you need to pivot it the other way You should be able to software link that I would guess but I'm not a software guy You don't really need to pivot on the radio access too much because you can't really conform to the shape of what you're running over anyway since your track units are too stiff You don't have anywhere near enough moving in the tracks along there track rolling access So you need to do is stop the dead track from getting stuck just have it pivot up so when you're moving forward that dead track would pivot so that you're basically riding on one edge of the rolling wheels on the edge of the track and then the portion of the track closest to the body would be angled up like 30° this way when it came into contact with something it would just push right over it

@florianfesti4325 Год назад

Minor nitpick: The groves on the wheels go the wrong way. Those little wheels work the opposite way than normal wheels, where you want traction when driving them. Here you need trac(k)tion when they are not rolling. So the profile needs to be turned 90°. As this is probably not great for 3D printing adding spiral grooves in both directions will also work.

@rakninja 11 месяцев назад

when i was in service, i was a cavalry scout on a bradley. this whole project has my brain screaming "HERESY!" i'm glad you came to the same conclusions i did that this would not scale up well at all. not to mention the problems with using such a configuration in the real world. see, the way tracked vehicles brake is track tension. with the wheels on the track, on unlevel ground, this design would not be able to stop. it would also have large control problems going up or down all but the most shallow of inclines.

@karyjas1 Год назад

If you took the rubber wheels within the tracks (at the bottom) and made them much thicker and with holes in them, like airless tires, you would also achieve a suspension effect, which tanks have. In fact, its so important for tanks that they have their own unique type of suspension. Also what youre trying to build vaguely resembles the omnidirectional treadmill for VR

@dillonhayward1870 Год назад

Imagine a real life military purpose triangle tank. What would its purposes be? This is just a thrown out guess but would a triangle tank be more efficient in an urban war environment. Because a triangle tank could turn extremely quickly enabling it to have better movement / agility property's. Also a triangle tank could still move about if one tred was to be blown off. Just putting it out there.

@jort93z Год назад

inb4 you get stuck on every curb cause of the tiny wheels. Don't think the tiny wheels are something that can be solved while still keeping the tank track shape.

@LordDragox412 Год назад

@@jort93z Just replace the tiny wheels with tiny tanks that have wheels replaced with tiny tanks that have wheels replaced with tiny tanks that have ( ... x infinity), problem solved!

@laz272727 Год назад

Tracked vehicles can already spin on a dime and smaller urban robots tend to have four tracks anyway, so this scheme is still worse than the default.

@xaytana Год назад

@@laz272727 This and these omni-track designs lose the benefit of effective surface area, the entire point of using a tracked vehicle. An ideal omni-track, which would essentially be a perpendicular track around the main linkages, would be overly complex and still have sacrifices. And grousers are impractical on omni-tracks due to the fact that you'd either need a segmented blade that can be rotated per perpendicular track, or you'd have to settle with using a studded design that sacrifices grouser effectiveness for simplicity; especially considering grousers will help with climbing capabilities, plus various capabilities in varying terrain types. Omni-tanks are impractical no matter how you look at them, and for additional maneuverability that really only replaces a one-point turn and a three-point repositioning, with some additional vector efficiency that's more than likely negligible at the end of the day; plus screw tanks exist for higher maneuverability, though their only downside is that they're less effective on hard surfaces, at least not without a segmented and positionable screw. There's just no point to omni-tanks, really the only practical use for an omni-track would be for something like an omni-directional walkable surface where you don't have to necessarily worry about a sprocket drive which simplifies linkage design greatly, and you don't have to worry about other factors other than keeping the surface fairly flat; which circles back to the optimal design being a perpendicular track around the main track's linkages. I really don't see this series evolving into proper design, and I really don't see it evolving into displaying the one practical use-case of an omni-track. Yet, as usual, 90% of the comment section doesn't seem to understand the basics of mechanisms nor have the logical thought to figure things out, and we all know James will never point these things out himself.

@nuclearmedicineman6270 Год назад

An agile tank still can't dodge an RPG fired down the street.

@andybrice2711 Год назад

Just put some low-friction ramps leading out from the centre, down to the underside of each track. Then the inside edges of the tracks won't get caught going up steps.

@IanMacKenzie_IIsWe 10 месяцев назад

A+ for design and build F for practicality

@xavierjiang7112 Год назад

Usually there's only one teethed wheel per track, the others are just smooth road wheels. There are alignment teeth on the track that stick out into the grooves of road wheels to keep it in the center. The tracks are really just to help spread out the weight of the vehicle. Omnidirectional Super Swamper!

@KarmaCadet Год назад

googley eyes always helps me keep track of where the things are facing

@alockpro Год назад

Prototyping with printers is great to see proof of concept. But don't forget scaling is important. The reason full-size tanks drive over things is in big part to the weight. But, as the weight gets bigger, things like small wheel parts could be too easily snapped off by the torque physics. Those police bomb robots climb stairs well, partly due to weight. If they were made to be super light, they might just slide down the stairs, flip over, or not even get started. I love the simple design of a three-wheel per axel, hand-truck dolly for climbing and driving.

@ImmortalAbsol Год назад

Could you put a skid/guide 90° to the track between the core and the track to lead the back track up to the obstacle?

@markfergerson2145 Год назад

You can make the wheels on the edges of the tracks larger without them interfering with each other by putting them on every other link rather than every link. The limit then will be the frames holding the sprocket and idler axles. The track frames could use spring loading to prevent them flipping over. This is such a bizarre cool project. I'm glad it came up in my feed. NASA and ESA need to hire you for planetary rover designing.

@SeaforgedArtifacts Год назад

This would make an impressive "howl's moving castle"

@diyfireking Год назад

Amazing job, I love your design.

@poshhippie6446 Год назад

Much larger idler wheels would probably help with the step issue

@thiccphone1166 Год назад

The DoD should hire you! I never could imagine someone could make the tank LESS EFFICIENT!

@jamason1235 Год назад

he is the mister maker of 3d printing

@wieneroach2280 Год назад

You should make the front be facing the long side of one track so that the back track doesnt get stuck onto whatever its climbing

@amirhm6459 Год назад

I didn't expect the track is struggling to move sideways, awesome build!!!!

@D3VASTATl0N Год назад

I really like seeing the iterations over time.

@Mysda_ Год назад

This sparked a question in my head that I cannot test. What would happen with Mecanum tracks? Only 2 parallel Mecanum tracks. Would it be able to turn? and how?

@Mysda_ Год назад

This reminds me of Screw-propelled vehicle, and I really wonders if something simple could be done for combat robots.

@THESocialJusticeWarrior Год назад

This a good example of how omni wheels only work well on a flat surface. Even if you make them into tank treads, they are even worse on uneven surfaces.

@jacobnassr-low7757 Год назад

To fix your issue regarding the obstacles, I would add a fourth track to make it a diamond tank instead of a triangle tank. Then hypothetically the two rear tracks should have just as easy a time getting over the obstacle as the front tracks.

@infinitelyexplosive4131 Год назад

Seems like this is one of those concepts that's just doomed to never work out. The perpendicular wheels are more or less fundamentally incompatible with purpose of tracks.

@MegaGaming11 Год назад

I feel like a way to deal with the Radius Issue might be to, instead of having the treads flat, have them flare upwards at the ends. That way you can make the radii of the small wheels bigger without having to touch the radius of the sprockets

@jort93z Год назад

i don't think thats really helpful. sure, the radius of the sprocket wouldn't change, but the radius of the track would. That's like simply making the track thicker.

@MegaGaming11 Год назад

@@jort93z Fair enough. There's also the problem that you'd have to make the track pieces longer to accomodate for the flare as well, so in a vacuum it sounds like a good idea, but actually implementing that solution would be a design nightmare. Part of me still wants to go with the "raise the small wheels so that they interact with the floor differently" route, but, as you stated, that *does* still increase the radius of the track. Doing that would also change how the tracks interact with the surface(s) they drive on, so it's not a solution that's feasible.

@jort93z Год назад

@@MegaGaming11 I mean, if you curve it up, essentially it would only drive on the small wheels, driving only on the rims. That sort of removes part of the benefits of the tracks in the first place. The benefit of tracks is that the load is spread over a large surface area.

@MegaGaming11 Год назад

@@jort93z True! a potential solution to that particular problem would to just be to extrude the pieces downwards to match with the now curved/extended part of the tracks, but that presents its own issues and also reintroduces the original issue of "Ya can't increase the radius of the small wheels without increasing the radius of the track" if it even got rid of it in the first place.

@jort93z Год назад

@@MegaGaming11 "would to just be to extrude the pieces downwards to match with the now curved/extended part of the tracks" ...And at that point you are back to basically just making thicker tracks. Which would let you increase the size of the small wheels.

@Sybrss Год назад

Hey, i used you model in a robotics competition at my school, thanks i got pretty high.

@dfgdfg_ Год назад

Woah, you're a smart and talented engineer!

@kameljoe21 Год назад

Super. I like anything track related.

@ylerta1 Год назад

An alternative to the hobby axels are 1.6mm TIG wire.

@MPee76 Год назад

Pivoting the tracks is a bad idea. That basically turns the third track into a plow that gets stuck on everything.

@AtomBomb420 Год назад

Hey James, an addition I would recommend would be to create an indicator of what direction is forward, the simplest I can think of is just a 3D printed part that just attaches to the front of the drone, like an arrow with high contrast, or it could be as fancy as you decide, I am no engineer, an I certainly know nothing about electronics, but if you were to want a fancy mechanism that shows your bearing, I should not elaborate, as I tried, but ended up scrapping that thought, knowing I would be horribly wrong as my inexperience shows

@johnnyhorton209 Год назад

Here is an idea for how to stop the back tree from getting stuck while traversing over the plank. Try putting a slope on the inside of the tread unit (not the moving treads) so that the tank will just slide over it.

@Puzzlers100 Год назад

I had a robot for vex in 8th grade that had omni-directional capability, and what I did to get over the issue of remembering where the front was was to convert the joystick input from X and Y to a vector, add the direction the robot was facing using a calibrated IMU, then converted that vector into information for each of the driving wheels. This meant the when I pushed the stick up, the robot always drove away, no matter which direction it was facing.

@madhukeshnp Год назад

As soon as he says, a quick ad from today's sponcer!! I quickly tap 6-8times on the right side of of my screen. QuickyTap - free ad blocker 😂😂😂

@jessewolf9145 Год назад

Great project! It seems like the whole driving experience could be improved by implementing driver-centric steering by adding a gyro and adjusting where forward is based on the gyro value.

@thewatersavior Год назад

#Sweet! Looks like you just need to flip the driving direction to keep the non moving wheel in front and push it over the obstacle

@dominantmale89 Год назад

If you could power the pivot point too, to rotate the track then you could raise the back of the unit to stop the 3rd track catching and also lift the whole body to stop it grounding.

@alifleming2256 Год назад

I'd be interested to see if you'd have more luck using only two tracks but rotating the rollers to 45 degrees like a track equivalent of a mecanum wheel. Potential a more usable result?

@AlRoderick Год назад

I'm not sure if that would work. I think you would need four tracks but I can't quite spell out why. I think it's because that in a meccanum wheel drive you need to be able to run the two wheels on one side in opposing directions, and I think any two wheels on the same edge of the vehicle have to have opposite chirality on the spiral wheels. With tracks effectively, the wheel pairs are locked together and have the same handedness. Also, the reason this works at all with the tracks is because each individual roller is capable of moving at a different speed, so it can turn. With angled wheels on the tracks most of them would be dragging anytime the tank would turn.

@josephpettit1519 Год назад

I think with some coding, you could impart a slow spin to the system while still moving in the direction requested. This would self regulate the rear track snag problem. However, you wouldn't know which direction is forward since 'forward' would be constantly rotating with the spin. You could mount a turret on top that always points 'forward'. And, if you go as far as a turret, you might add a camera so that you get a robot POV. The flipping track issue might be solved with some pins and rubber bands so that as track starts to flip, the rubber bands start producing more and more counter force to bring it back to neutral/level.

@androswolf Год назад

You totaly need to combine this bot with that head-bot you mentioned !

@jamesfreach1022 Год назад

This is fantastic 👏 👌 If you can set it up to run G-code and attach a wood router you'd have a cheaper version of the Goliath C N C Machine which costs about $8000.00 US Dollars. The low profile track isn't a mistake it just needs to be utilized in the correct project like a clone of Goliath. The router only needs to be mounted on a Z Axis Slide and the tracks controlled similarly to the Mazlow C N C Machine

@renaion Год назад

man said good at getting over things they are huge gun on tracks

@jacobgoldstein6878 Год назад

Making the swivel axis motorized could mean you can lift up that back track when it gets stuck which could help with clearance

@widgget Год назад

i was thinking this while watching. would make for some interesting looking movements like being able to stand up on the tips of the tracks.

@wurstelei1356 Год назад

The drive mode where 2 tracks move half speed of the 3rd track is already very good at climbing things but your idea still is very smart. The angle of the tracks on ground would change with your idea thou. It also would work if you rotate the tracks 45 degrees around their long side so the tracks form ramps to the sides. You have to take obstacles driving backwards in this case.

@TheNolok3428 Год назад

Very cool. Outdoor Test would be nice.

@wisselinnemans6190 Год назад

To solve the direction issue you could use a gyroscope and use that to make a north-east-south-west and drive based on that instead of the directions of the tank. Basically using the orientation to the room/starting position to decide which way to go. Problem is that you would have to stay on the same side/place in the room or relative to the starting position.

@chaselewis6534 Год назад

Seems like there needs to be a suspension system of some form so the front 2 tracks can pull the back one over. The issue seems to be from the rigidity of the mounting than any driving issue. Overall cool build.

@Nightelfmusic Год назад

30 degree angle on the edge of each track plate. Three wheels on each track plate wheels closer together almost touching. Should help the next wheel bump up. And fit a servo to force flatten each track module to horizon slowly . So it can Force the tank to climb and get a different foor print

@ToninFightsEntropy Год назад

Or 3 tracks pointing out like a star, but one able to rotate 90 degrees on the yaw?

@vernonzehr Год назад

quite frankly not sure why this is better than a two track tank type movement. the three tracks always all of them all the time are fighting and going against the direction they are designed to move, the 120 degree angle is always going to be “wrong” with wasted effort/energy. than add in the third track that catches on stuff. the only benefit is a tight turning radius but both can spin in place. driving over obstacles requires a tremendous amount of fiddling extra movement. it seems pointless to me but what do i know.

@207shark Год назад

What a redesign!

@Darkerfoxtech Год назад

"If there are only 4 directions why are there 6 pedals" in real life. Never thought I'd see the day.

@444haluk Год назад

You know what would be great: make it tetrahedral!

@BjarkeHellden Год назад

I suggest making 4 tracks tracks in a square configuration and have the tanks move diagonally as it's forward and sideway directions. This way it would not get stuck.

@g80gzt Год назад

Maybe a method to tilt a track, so the inward tilts up and the outside tilts down, to help angle it over when it gets stuck as prevalently shown.

@the_jingo Год назад

Cool concept but would’ve never work irl for so many reason biggest one being maintenance

@thebclpoonilli3098 Год назад

tracks in the tracks, instead of wheels. it will give more even traction. the inner tracks should be the only part touching the ground. the inner tracks can be thick rubber bands on pullies. fun video, thanks for the ideas. what are done with old models?

@Adierit Год назад

Have you tried track extensions that are angled? it would solve the issue of the rear getting caught, or replacing the wheels with rubber ones that have compliance so they can deform to grab the surface better.

@VerdASMR Год назад

7:57 Do you have any references for why this is the case? I think I'm pretty okay at trig but im not really sure how you've applied this/ how you intuit this

@liambohl Год назад

James's trig is incorrect, which is understandable because omnidirectional wheels and treads are really tricky. When the tank moves in the direction of one tread, that tread should move at full speed, and the other two should move at half speed.

@liambohl Год назад

Understanding this requires some vector arithmetic. If the tank is not rotating, then the velocity of the tank equals -1 * (velocity of the bottom surface of a given tread relative to the tank + velocity of a small wheel relative to that tread). Since the latter two velocities are perpendicular, the vector sum can be drawn as a right triangle. Given the velocity (speed and direction) of the tank, you can use this right triangle to compute the required velocity of the tread.

@JosephMarinier Год назад

Same observation. I think Liam nailed it: the multiplier in the code should have been 0.5 instead of 0.66.

@CryptomasterLeviathan Год назад

Make a 34 little powered tracks and put them all perpendicular on the bigger track! And make three of them. That would be cool. Haha, keep making amazing videos!

@Thomo27 Год назад

if you had a large weight in the middle, would you be able to add a mechanism to lift one track an inch or two. then if it realises one track is stuck, it can just lift it up slightly and depend on the other two

@theredstormer8078 Год назад

I think it would be funny if instead of tiny wheels you used little tiny rubber tank tracks. It would take forever to assemble and probably have way too much friction, but it would be funny.

@hamjudo Год назад

True *omnitrack*

@haassteambraker9959 Год назад

I wonder how practical it'd be to replace the omni rollers with more tracks, using bands instead of links for obvious reasons. Could get edge grip and have a trapezoid or otherwise convex shaped track links to better enable the tailing track to climb obstacles.

@carlisackson7490 Год назад

An alternative to making the little wheels bigger is to either angle the tracks toward the outside edge lifting the inside off the ground or making the tank level itself so that it lifts the back track up to the level of the higher surface

@RandoWisLuL Год назад

2 more improvements: Make it so the tracks can't swivel too far and make it so the tracks can swivel up as one piece that way you don't have to make the wheels bigger.

@IraQNid 10 месяцев назад

a life size version should be made to test how well it really performs as the basis for a drive train that we can ride on. Three treads could mean no need for insurance, driver's license, or other government interference in our lives.

@bbgun061 Год назад

I suggest adding springs to the articulated tracks so they are forced back down onto the surface when going over obstacles. You probably won't make a third version, but for future projects that's my 2¢.

@timseguine2 Год назад

Maybe a wye configuration would work better than a delta configuration like you have here? I think that would avoid the problem of the back track getting stuck perpendicular.

@robbydonaghy8735 Год назад

I was thinking that too, but how would you turn? All the tracks would point to the center of gravity. And I think even moving just 2 in opposite directions would make you mostly strafe.

@timseguine2 Год назад

@@robbydonaghy8735 Ah, yeah, no way to generate torque to rotate. That's right. I was only thinking about the one problem and created a worse one.

@ThrunRC Год назад

Great idea 👏👏👏

@hakanloyan3483 Год назад

nice and clear transition. Good job bro.

@amppari_234 Год назад

An insignificant correction, but the part you call the idler in this case is not an idler, but a roadwheel. An idler is a pulley that transmits no power but guides or stretches a belt or rope. And since the track already goes there without the roadwheel/idler, it is indeed just a roadwheel. And a roadwheel on the other hand is a part of a tracked vehicle that supports its weight. But a cool creation anyway.

@Blackwind_Legacy Год назад

My only suggest is to go to 4 treads now. Raise the2 not in use. All 4 down for a fast spin. 2 down for traditional movement. Change which set is down for sideways movement.

@handy-capoutdoors4063 Год назад

It is very interesting concept. I really like the ability to crab walk without turning the chassis. However it does seem like a design with high ware on the mini wheels. And very little ability to traverse small changes in terrain. I could see this robot base design in use in warehouse facilities with waxed smooth floors. With a "pick and place" manipulator. Or in a mail sorting room in an office building where there are multiple standing shelves. And few if any door thresholds or cracks in the floor to Traverse. Basically adding a 3rd dimension to the current vertical/horizontal rail mounted pick and place systems. The downside aside from issues with small elevation changes would be noise and weight compacity. The wheel matereal would play a huge part in the rate of ware but it would also ware on the floor surface. The amount of ware would depend on the weight being put on the base and is somewhat proportional to the amount of noise. Have you thought about using screws instead of tracks or wheels? I'm curious if screws may reduce ware on the mini wheels. The idea of splitting the diffrence in angle. The deeper groove of a screw may give you better traction when encountering sudden changes of terrain and the curve of the screw parallel to the edge of terrain could just roll over or assist. Just some thoughts.

@kaigraham1287 Год назад

wider beveled tracks? maybe two or more tilled tracks , so the tracks contact the obstacle at an angle or you could cut the track in half and make it Three-pointed star shaped or other star shapes? maybe the triangle configuration only really works for wheels

@darth_dan8886 Год назад

These pivots most certainly need some spring or rubber powered suspension to get better traction upon lifting itself onto surfaces. I also wonder if it'd help any if the tracks had a slight \_/ shape with the wheels positioned in the corners. Now that could potentially become an odd shape and snag on things... But I frankly am not sure how to solve the back track perpendicularity issue. Lift the whole track up mechanically? 😅 Guess there's a reason we have omniwheels and not omnitracks - people haven't figured out how to solve all these little problems yet. Also I would suggest adding a directional sensor like a compass or a gyroscope, and correcting the drive vector to that direction rather than following a fixed forward configuration. It'd be much more intuitive to control, I think. And would also open up a possibility of a "spin and forward" maneuver.

@gregoetker1389 Год назад

Try doing it with a bulldozer track design which would allow two larger diameter sprockets at the ends. That would greatly reduce friction. If you want a slight tilt up at each end look up a T34 tank.

@LtColShingSides Год назад

RU-vid recommendations came through, for once! lol This was really neat!

@jackyzhou3642 Год назад

Quite an impressive project. While I have absolutely no idea how to build a robot, I'm leaving a like for keeping it open source. Good job mate!

@blenderbuch Год назад

Rotate constantly and overlay the absolute direction of movement 😉

@benfuhler9556 Год назад

Would it be possible to make the robot constantly spin that way the obstacles would be easier for the “driver”?

@VentureSoftCo Год назад

It's crazy how it only took you 12 minutes to build this!

@GordonWrigley Год назад

You could put a rotating head on it and remap the controls so forward is the direction the head is facing and you get controls for turning the head or the body while keeping the other one mostly still

@scottsamborn4639 Год назад

With all of the stuff that gets 3D printed, I have to wonder if there is a storage unit somewhere chockablock with PLA or if you are recycling much of it.

@joshuascreatives2281 Год назад

Instead of painting and adding little wheels ,, you can just add small or average size metal balls,, it might solve the issue of getting stuck while climbing up the obstacles

@bobverhoeven4149 Год назад

Two track with omni wheels simpeler and better i think.Superb channel.

@ОлександрРибалко-й6т Год назад

Thanks for doing such nice things! What about idea: tank join two things tracks and small wheels. so its ability limited to small whells. Then, why do you need tracks? Thanks!

@TheNynrahGhostJo Год назад

What if the tiny wheels were little treads instead?

@H34... Год назад

I think the big issue you cant really overcome is that trapezoidal tracks only help when the track is going head on. When collides side on, the edge of the track being flat on the ground can't really go over anything. Still cool.

@TalynStormcrow Год назад

This can be fixed by keeping the tank always slightly rotating.

@locouk Год назад

You need to replace the small wheels with TPU tracks, do you use silicone spray to reduce friction?

@---l--- Год назад

Beautiful

@RuthlessMojo Год назад

Imagine what would happen if this side moving track design was employed by modern military vehicles. It would allow for unprecedented manoeuvrability. Add side thrusters and you could rapidly side dodge.

@odman7945 Год назад

Feel like if you have to do a lot of pivoting to get over obstaclea one might as well just use 2-tracks and pivot to get over easily. So some solution to the backend track might be needed.