[EV3] Omnidirectional Robot Concept and Discussion

[Alasdair Ryan]

How feasible would it be to make something that could climb stairs?

Making something that travels on flat ground will not need suspension,but one idea if you wanted suspension is not to use springs but make each axle "float".

Stair Climbing Robot

Edited August 10, 2016 by Alasdair Ryan

[Jim]

A stair climbing robot is very cool, but that's not the intention of this robot. Since it will be quite high, it will fall off the stairs easily.

It will indeed mainly travel on a flat surface. Having suspension won't add much significant value.

[Alasdair Ryan]

But if you used a tilt sensor and some motors to level the body of the robot,it might work.

[Jim]

I will try to finish my initial concept. That will be difficult enough, given my track record of finishing projects

[Alasdair Ryan]

Ok,no you should not suspension.Would I be right in thinking if you build a four sided robot then it would be harder to steer than a three sided robot,as the wheels are at 90 degrees?

[Jim]

Depends on the setup of the four wheels. Two on each side and one in the front and one in the back (+ shape). Or in a X shape.

I think when using an + shape it won't be too difficult to drive, since you basically have a two wheeled robot with two casters.

[oracid]

How feasible would it be to make something that could climb stairs?

The goal of Jim is very different, but if you're interested, I invite you to look at this: http://www.eurobricks.com/forum/index.php?showtopic=119844&hl=climber

The suspension for a robot, it's a real discussion !

I think if the goal is to consume less energy, it can be interesting, but if set for user comfort as in a car, and there, there is no user...

[Jim]

The suspension for a robot, it's a real discussion !

Ultimately it would result in the robot less shaking when it drives over uneven surface. But with a robot like mine....that probably won't happen. I'm thinking more exhibition hall floors, which tend to be flat.

[oracid]

Ultimately it would result in the robot less shaking when it drives over uneven surface. But with a robot like mine....that probably won't happen. I'm thinking more exhibition hall floors, which tend to be flat.

I fully agree with you!

[TheMindGarage]

How feasible would it be to make something that could climb stairs?

Making something that travels on flat ground will not need suspension,but one idea if you wanted suspension is not to use springs but make each axle "float".

Stair Climbing Robot

I have made a stair-climbing robot, but it was very difficult. I had three axles fitted with independent torsion-bar suspension. Each axle was able to slide up or down (central differential and all) up to 8 studs using the 42043's outrigger parts and worm gears (that setup is extremely strong and has survived a load of about 18kg per slider in a different creation). I managed to develop a compact, high-torque Schmidt coupling to link the axles' driveshafts to each other and to the gearbox. Gearbox had 7 positions - four were forward speeds and the other three controlled the sliders. So to climb a step, the gearbox goes into neutral and engages the first slider, then the drive motor raises the first axle. Then the vehicle drives forwards a little and repeats this for the other two axles. The suspension and gearbox were really pushed to their limits.

[Jim]

Here's an example of a 4 wheel setup:

[Alasdair Ryan]

They are better football players than the Dutch national team.

For your robot I would continue with the idea you had already and build four "cubes" each with a motor and wheel,then form a bigger square or triangle with the way you connect all of the smaller "cubes" together.

Edited August 16, 2016 by Alasdair Ryan

[Jim]

They are better football players than the Dutch national team.

Can't argue with that

For your robot I would continue with the idea you had already and build four "cubes" each with a motor and wheel,then form a bigger square or triangle with the way you connect all of the smaller "cubes" together.

Exactly! I'm working on one of the cubes as we speak. I will post my progress later today.

Funny thing is, that while I have pretty big collection, I still seem to be missing quite few simple parts in large numbers.

[Jim]

I have tried both LXF's and Alasdair's designs for the wheel hubs. However, they had too much slack. So I invested some time to come up with a new design. This design proves to be simply, yet very efficient.

I have used two 2x1 Thin Liftarms to connect the axle. This proves to be very sturdy, without too much slack. Hail to the new 3L Axle with Stop.

After designing the new wheel hubs, I continued with the design for the entire wheel section.

As you can see the left side is slightly different than the right side (which is one stud lower). I will choose which design is better when I finish designing the sides.

The sides still need a lot of work. Not sure on how to design them.

Here's a bottom view from the inner side.

Obviously, I need to implement some proper bracing, but overall I think this is a design I can continue working on.

I'm growing very fond of this new part by the way. It almost feels like cheating when you're using this part, it's so convenient.

[Lego2016Lego]

@Jim

Good job and I love that part too!

Edited August 21, 2016 by Lego2016Lego

[JJ2]

Very nice though that is a 3x1 thin lift arm not a 2x1, also that part is one of the best every order I do I get more of them if I can, I always end run out or them.

[syclone]

Very nice though that is a 3x1 thin lift arm not a 2x1, also that part is one of the best every order I do I get more of them if I can, I always end run out or them.

You should look better at the photo ;) (Note the half bushes)

@Jim I would say almost hail to the 3l axle. It would be more convenient in other color (LBG, DBG or black)

Edited August 21, 2016 by LXF

[Jim]

You should look better at the photo ;) (Note the half bushes)

Well spotted. The 3L is indeed of no use for the main axle.

@Jim I would say almost hail to the 3l axle. It would be more convenient in other color (LBG, DBG or black)

True, but that goes for other axles as well. Same for the 5L with Stop. I'm glad they are available to begin with.

[Jim]

I'm trying to finish one of the wheel hubs. The goal is to make it rather closed design, so not too many gaps visible. I'm particularly struggling with the corners (obviously). I'm open to suggestions for an elegant connection between the two bent liftarms.

I think this is a particular cool design, and I will more or less try to replicate the color scheme.

(however, the color scheme isn't important right now. The actual design is)

Here's the updated file:

https://www.dropbox.com/s/gf4ag38bvdt0tlz/Base-v2.lxf?dl=0

[Jim]

In other words...how to design the area between two bent liftarms in an elegant way?!

https://www.dropbox.com/s/mvopicqrtnqqe4i/BentLiftarms.lxf?dl=0

[Charbel]

How about this Jim ?

[Aswin]

Hi,

Writing a program to control a holonomic robot can be a pain. If you do not want to go down that path yourself you can consider using leJOS. It has classes to control a holonomic robot with three or more wheels. There is no limit to the maximum number of wheels so it fits a modular design very well.

Using this class is really easy. You describe the robot once: what is the placement of the wheels, the wheel size, gearing etc. Then all you have to do is to instruct the robot to move: specify forward speed, direction, and rotational speed. The class then takes care of everything else, even ramping up or down and the smooth transition from one move to another.

I developed this class for my holonomic robots. It even works for my holonomic walker: https://www.facebook.com/groups/legomindstorms/permalink/678340108980885/

More info about the class can be found here: https://lejosnews.wordpress.com/2015/05/12/lejos-navigation-the-chassis/

Edited August 25, 2016 by Aswin

[Jim]

Sweet, that's really useful information. I was planning on doing RobotC and programming is part of the fun. But I will consider leJOS. Actually I need to dive in all the current options for the EV3. I used to be well informed in the time of NXT, but I haven't been investing much time in EV3.

Your walker is very cool. When I saw the video I somehow felt it was your creation, and it indeed was. Care to make a topic about it here?

Using this class is really easy. You describe the robot once: what is the placement of the wheels, the wheel size, gearing etc. Then all you have to do is to instruct the robot to move: specify forward speed, direction, and rotational speed. The class then takes care of everything else, even ramping up or down and the smooth transition from one move to another.

Sounds awesome! Maybe I can port it to RobotC.

I developed this class for my holonomic robots.

You developed it yourself, cool. Now I feel obliged to try it

[Jim]

I noticed something odd. Apparently, this doesn't fly?!

I did a quick test in real life and I thought it would work, but "computer says no!".

Edit: actual test confirmed that it's not a valid connection.

Is there a source where I can find the legitimate combinations? Sariel's book maybe?

How about this Jim ?

Sorry for the late reply. I totally missed your post. Seems like a nice suggestion, thanks!!

I'm quite happy with my current setup, but I will give it go

[skppo]

Yeah there are no integer solutions to the Pythagorean theorem (a²+b²=c²) with where a=b.

Here are a few useful ones though

3 4 5

6 8 10

5 12 13

9 12 15

8 15 17

Just remember to count the distance between the centers of pin/axle holes not the length of the element.

(The first one can be built with liftarms of lengths 4, 5 and 6)

Edited September 6, 2016 by skppo