Moving multiple objects independently on a flat surface

Thread Starter

Maskawisewin

Joined Oct 7, 2019
7
Is there a standard method of moving multiple objects in any independent direction? Their position would have to be known and guided almost constantly. They need to move on a flat plane.

I can take almost any approach but the smaller the objects the better.

So far I've come up with/seen:
  • a gang of cylindrically shaped micro robots. The robots can be tall, but not wide. I was looking at TinyCircuits to prototype these. Problem is they're not as precise as I would like and quadrature encoders take up a lot of space based on my experience. Plus the motors would have to be vertical and gear at 90 degrees. It seems possible, but a huge headache. Plus if they're too tall, they could fall over if they collide.
  • something involving multiple rows and columns of belts and pulleys that move the objects. The issues I see with this is each object needs to move independently and there would be collisions.
  • magnets above the plane that are moved from below the plane... somehow. I feel like there would be tangled wires with this solution.
  • a matrix of solenoids. This isn't moving the objects, but it would accomplish the same thing. Problem with this one is the solenoids themselves would take up a lot of space and not provide much "physical resolution" per say. Maybe my assumption is wrong.
I hope there might be an existing industrial solution for this problem and I just haven't seen it.
 

ericgibbs

Joined Jan 29, 2010
18,766
hi M,
Welcome to AAC.
Some additional information will be helpful.
What are the objects , how many and the size and material of the flat area.

What is the purpose/application of this project.?
E
 

Thread Starter

Maskawisewin

Joined Oct 7, 2019
7
hi M,
Welcome to AAC.
Some additional information will be helpful.
What are the objects , how many and the size and material of the flat area.

What is the purpose/application of this project.?
E
The objects can be anything really. I'm trying to communicate basic positional information across a plane. So the plane can be any material. It can even be electrified if needed. I'd say I'd need anywhere from 13 to 23 objects on the plane in different areas. But once the number of objects is decided, it stays constant. The size of the plane can be any size, but the smaller the better.

This is why my initial idea was an array of solenoids but I did not feel the movements would be fluidic enough.

Basically, if a 3d tactile display existed a basic app would negate the need for this project but that doesn't exist yet, so I'm trying to create something far more basic and without any display technology.
 
Last edited by a moderator:

Sensacell

Joined Jun 19, 2012
3,432
I situations like this, the solution lies within the specifics.
You need to narrow this down substantially.

Size?
Weight?
Distance?
Speed?
Accuracy?

Define what you need to do exactly.
 

Alec_t

Joined Sep 17, 2013
14,280
If the object were spherical I think you could have a couple of internal motors each acting as a pendulum bob in a gimbal mount and driving in orthogonal directions. The object would follow the moving centre of gravity. Collisions between objects could occur if the object movement is autonomous rather than guided, and should be allowable if the motions are truly independent.
 

Thread Starter

Maskawisewin

Joined Oct 7, 2019
7
I situations like this, the solution lies within the specifics.
You need to narrow this down substantially.

Size?
Weight?
Distance?
Speed?
Accuracy?

Define what you need to do exactly.
Size: As small as possible while still visible to the naked eye (so basically as small as possible). I'd say the largest it could be is about a 5 cm diameter but that is pushing it. If I can get a 1x1 metre plane, then I'd hope to get 1 cm diameter but that's likely a pipe dream.
Weight: Likely guided by required speed. Can be as heavy as needed as object will not need to be lifted.
Distance: Ideally no more than 1x1 metre, but could be more to start with if that's all that is possible.
Speed: From 0 to 0.x m/s where x is the size of the plane.
Accuracy: Error can be quite high based on input. For example if the input from two objects are overlapping, they can be next to one another.

If the object were spherical I think you could have a couple of internal motors each acting as a pendulum bob in a gimbal mount and driving in orthogonal directions. The object would follow the moving centre of gravity. Collisions between objects could occur if the object movement is autonomous rather than guided, and should be allowable if the motions are truly independent.
How would the objects separate from one another once they collide? Each object (up to 23) would need its own centre of gravity.
 
Last edited:

atferrari

Joined Jan 6, 2004
4,764
The displacement/s of each piece, would they follow a fixed-already pattern or could it be changed in the future? That precludes anything mechanical with fixed dimensions, I think.
 

Thread Starter

Maskawisewin

Joined Oct 7, 2019
7
One motor could auto-reverse.
Ah okay, well then I'm not correctly visualizing what the original suggestion was. Will do some googling to see if I can piece it together.

The displacement/s of each piece, would they follow a fixed-already pattern or could it be changed in the future? That precludes anything mechanical with fixed dimensions, I think.
The pattern is random and not predicable and the speed of each object is always changing. Every object would be fixed in dimension. I think that answers your question but if not please elaborate.
 

BobTPH

Joined Jun 5, 2013
8,804
Do they all move at the same time, or is it okay if only one can be moving at a time?

(this is beginning to sound like a game of 20 questions.)

Bob
 

Thread Starter

Maskawisewin

Joined Oct 7, 2019
7
Do they all move at the same time, or is it okay if only one can be moving at a time?

(this is beginning to sound like a game of 20 questions.)

Bob
Apologies, and yes you're right it is.

A good analogy is if I was taking the positional data of insects and replicating thier movement. There is no need to worry about the input; assume it's constant and valid at all times. The bugs are going to move at different speeds, some will stop moving. Some will stop to smell or communicate with one another. However the replicated movement should move in unison with one another. If a bug is humping another bug, those two bugs can be next to one another. If the positioning gets a bit screwed up, for example one bug went around another and their position is off, then that bug will have to speed up to sync back up with the baseline time index.
 

BobTPH

Joined Jun 5, 2013
8,804
Okay, that eliminates the suggestion I was going to make.

This is a very difficult problem. I see no way to avoid having each object be a self contained robot.

Bob
 

Alec_t

Joined Sep 17, 2013
14,280
Rather than actually build these bugs, couldn't you just simulate them on a PC? Their semi-random motion would be a doddle to program.
 

djsfantasi

Joined Apr 11, 2010
9,156
Could the objects not actually move, but be represented by a graphic pattern?

In that case, your plane could consist of a number of pixels. The pixels would be divided into a number of positions or cells (cells within your grid).

The cells could be a matrix or grouping of surface mouth LEDs and be addressable.

Then, it becomes a graphic display problem. Your (microprocessor) control would sweep over all cells, setting the cell to a specific object pattern/color.

Advanced programming could simulate the actual movement, both in direction and speed by animating pixels across cells.
 

WBahn

Joined Mar 31, 2012
29,976
I wonder if you could do something like a 2-D linear motor and then have it finely enough segmented so that you could drive only the portions of it that are under a particular object at a given time.
 

Thread Starter

Maskawisewin

Joined Oct 7, 2019
7
Unfortunately the main stipulation of this particular project is for it to be tactile without user intervention (like a pen input). If it was just screen output, I would have been finished by now.

I wonder if you could do something like a 2-D linear motor and then have it finely enough segmented so that you could drive only the portions of it that are under a particular object at a given time.
What happens when one object circles another and both are moving at the same time? If each object can move at any time, and are all likely constantly moving all at the same time, then I don't think that will work.

I'm really starting to think a grid of solenoids or a technology like rods of a shape-memory alloy may be best at this point.
 

WBahn

Joined Mar 31, 2012
29,976
Unfortunately the main stipulation of this particular project is for it to be tactile without user intervention (like a pen input). If it was just screen output, I would have been finished by now.



What happens when one object circles another and both are moving at the same time? If each object can move at any time, and are all likely constantly moving all at the same time, then I don't think that will work.

I'm really starting to think a grid of solenoids or a technology like rods of a shape-memory alloy may be best at this point.
As with many things, you can create the illusion that they are moving at the same time. Move A a small amount the move B a small amount and then move A a small amount and if you do this fast enough, they will appear to both be moving simultaneously.
 
Top