# Traction mechanisms for a robot in confined spaces

I'd like to build an exploratory robot that can fit within confined spaces1.

What traction mechanisms should I consider to propel a robot along the inside of a 4" pipe?

The pipe to be traversed will likely have fluid (assume water) within it to varying degrees from partially full to completely full. I'm initially planning on tethering the robot with coaxial cable.

So far, I have considered a 3 and 4 point wheeled approach, such as with a Y or X configuration of wheels. And I have considered treads instead of wheels. The robot will need to be able to navigate 90o turns on the way in, but can be hauled out via the coaxial cable if need be. Preferably, it will be able to back its way out for the majority of the distance it can cover.

The robot would not need to climb unassisted. The initial use will be in a pipe that drops several feet from the access point, and then slopes down approximately 5 feet over a 60 foot run. Retrieving the robot will be assisted by the coaxial cable or a tether line.

1 It's a rather boring application, actually. I want to see how badly the roots are creeping into the sewer line from my house.

• It doesn't address your question directly, but if there is some pressure in the pipe, this is typically done with a practice called 'pigging' - en.wikipedia.org/wiki/Pigging If you need to do it independent of the liquid, I would look at systems that have two opposing feet like tunnel digging machines. Feb 3 '15 at 5:45
• I was about to write an answer about sewer robots, but maybe you looked at esisting designs allready? Will the robot need to climb? Can you expand on the research you did so we know what we don'T needto tell you?
– mart
Feb 3 '15 at 7:59
• I'm really tempted to suggest some BristleBot variant.
– SF.
Feb 3 '15 at 11:49
• @mart - I haven't researched sewer bots at all yet. Based upon my conversations with various plumbers, they haven't heard of anything that does what I'm thinking of. I updated my question to provide a bit more context about what I'm trying to accomplish.
– user16
Feb 3 '15 at 16:17
• See also Figure 7 in this document das.ufsc.br/~vinicius/projetos/dutos/00257076.pdf. They speak about clamp & pull, i.e. where the robot can push against the pipe at the front and the back and is stretching/contracting in the middle. Feb 3 '15 at 16:23

First, I'll say that from personal experience, locating and implementing treads on a robot that size can be VERY difficult. I was building a robot that was supposed to be 8"x8"x20" (WxHxL), including the treads, and we had a lot of trouble finding usable treads that would be small enough but still hold up to some more intense use than Lego or K'nex robots. If you can find treads, you still need to be able to tension them and prevent them from traveling axially along the guide wheels (i.e. slipping off while the robot is driving.) So my first recommendation is to go with wheels.

As far as getting traction I would consider some sort of rudimentary suspension system, or inflated or soft rubber tires that have some give in them. Basically, something that allows you to design the robot to have a resting diameter of 4.5", but then compress to 4" to fit inside the pipe, put a higher normal force on each wheel, and increase the friction between the tire and the pipe. (Those numbers are examples, I'd do some more work to get an actual determination of ID, 4" probably is the OD, and to determine how much flex you can have. And be careful, too much friction will prevent the robot from actually moving.)

Other thoughts: I like the Y configuration the best, and you may actually want the top wheel to be smaller, just to stabilize the robot, and not provide any driving power through it. Consider how you're going to get power to the wheels if they're not on the same axis. You may have to power each direction of wheels independently, or your gearing may get complex. And you may not be able to offset the power transmission primarily off the axis, there may be radial translation as well.

You are looking to build a sewer robot, so you maybe want to look at other sewer robots. I have no diret experience with designing or handling sewer robots, but I can talk about the ones I saw at trade fairs.
Some builds have sets of 3 or 4 wheels with a mechanism that pushes the weels against the pipe walls, or - more typically - sets of 2 wheels. In the first case, the wheels are quite small. In the latter case, the wheels are sometimes shaped so that when looking at the robot from the front the cross section is more or less circular. Find a handful of pictures of the seconfd type here. The first type, with an X shape, you can see here. I think it's far larger than 4". Not that this is not a classical sewer robot but for some pipe building application that I don't understand.

One small robot I have seen had a small pivotable arm at the front with a light source and a camera. At a branching in the pipes, the robot could push the arm into one of the branches. When the robot moved forward, it would be slide into this branch because of the arm.
I've never seen tracked robots that would fit into a 4" pipe - again, don't take my word for it.

Note Most hits for the term sewer robot where on german sites. Maybe the actual english term is different, and only germans insist on calling them sewer robots.

• The "big reveal" on what I'm hoping to do is in my other, related question.
– user16
Feb 3 '15 at 21:00