Society of Robots - Robot Forum

Mechanics and Construction => Mechanics and Construction => Topic started by: dustynesmith on June 21, 2011, 09:58:09 PM

Title: Need Help To Design Roofing Robot
Post by: dustynesmith on June 21, 2011, 09:58:09 PM
I suffer from an illness that may cost me my job if I don't find a solution soon, and I'm hoping technology may be the answer. I have to get up on roofs to examine damages. I am interested in purchasing or designing a robot to help me with the roof work while I stay at ground level controlling it and viewing the damages from the robot's perspective. If anybody has any information that could help me out it would be greatly appreciated, and maybe a test for a new product. I'll give a list of things to consider if something like this does not yet exist, as I have not run across one in my career.

1. It must be a mobile remote control device that is able to initiate small tasks while moving on sloped or flat surfaces and negotiate valleys.
2. The main tool to consider is a quality camera (around 10 megapixels at least) that can preferably be seen from the ground on a laptop. Other tools optional are a pitch gauge, and shingle gauge, and an arm that can use sidewalk chalk to circle damages.
3. It must be able to stay on high slopes without trouble, so the wheel type is important. Ability to handle at least a 10 pitch roof is preferable.
4. Flight is optional, or a mix of flight and ground mobility.
5. Reasonable pricing is preferable.
6. Any way to view what the robot is doing from the ground is preferable.
7. It needs to take measurements of the roof
8. It should withstand hot summer heat and icy cold winters.

If you know of any way to help, please let me know.
Title: Re: Need Help To Design Roofing Robot
Post by: dustynesmith on June 21, 2011, 10:15:48 PM
...
Title: Re: Need Help To Design Roofing Robot
Post by: garrettg84 on June 22, 2011, 05:31:07 AM
I think this is a good idea. I could see this being accomplished with a simple low profile track driven bot, wired or wireless, with a simple 'probe' arm and a servo mounted camera setup. Pitch can be handled with a gyro. The only issue I could see is getting over the peak or dip of a roof. You might have 4 sets of tracks instead of just two and allow those tracks to angle themselves to stay in contact with the surface. I'm not entirely sure I understand 10 pitch, but assuming my googling was fruitful a 10 pitch roof sits at a 40 degree incline. I believe a low profile tracked bot could handle pitches higher than this provided the surface was solid and intact.

One other element for concern might be bot weight. How do you envision getting this bot to the roof? Is flight necessary? Flying tends to complicate things drastically when there are obstacles  present such as trees or chimneys.

From the looks of your post, you want somebody to build this for you, is that correct?
Title: Re: Need Help To Design Roofing Robot
Post by: waltr on June 22, 2011, 06:53:28 AM
I'm seeing a Quad-copter like vehicle that can hover at any point over the roof with an on-board wireless camera.
 
Title: Re: Need Help To Design Roofing Robot
Post by: garrettg84 on June 22, 2011, 07:37:54 AM
I'm seeing a Quad-copter like vehicle that can hover at any point over the roof with an on-board wireless camera.
 

I had originally thought the same thing, but I am in the process of buying a home and just did a roof inspection with my home inspector. One thing mentioned in dustynesmith's post was chalking the roof. I don't know what kind of quadcopter would have the stability to chalk a roof around a problem. The other issue is that under some circumstances my inspector used a 'probe' (glorified putty knife with an extended handle) to lift and check under some of the shingles.

Now if the quadcopter could be made to land on the odd pitches do its marking/lifting and then take off again I think you'd have a success. The only other issue, most video/images I've seen have the blur from the vibration of the brushless motors/props. While a SUPER fast snapping camera and image processing could also counter act this, I'd say it would start to cost a bit more than this gentleman might be interested in. The only other issue I can see is weight vs flight time. Add a camera, chalk, a probe and all of the electronics, batteries, motors, and servos to control it you end up with something that actually has to be substantial in size.

If a quadcopter is the way to go though, hobbyking has a quadcopter frame for about $15 that already comes with a servo operated/stabilized camera mount if you provide the servos and gyros for the stabilization.
http://www.hobbyking.com/hobbyking/store/uh_viewItem.asp?idProduct=16583 (http://www.hobbyking.com/hobbyking/store/uh_viewItem.asp?idProduct=16583)
Title: Re: Need Help To Design Roofing Robot
Post by: dustynesmith on June 22, 2011, 11:04:25 AM
I'm willing to discuss having somebody build the bot for me within a reasonable pricing for materials and labor if you have a portfolio and some references. I may consider doing it myself if the cost gets too high.

As far as getting up on the roof, either one man should be able to carry the bot with one arm up a ladder. If this is not possible I think hover flight becomes necessary.  At that point weight and battery life also become a bigger issue. I would suggest that it stay on the ground through some testing, then maybe upgrading to flight at a later time. You can expect this thing to put in lots of hours of labor and need to have the fortitude to come out ready for another day.

Thanks for the input.
Title: Re: Need Help To Design Roofing Robot
Post by: joe61 on June 22, 2011, 11:41:27 AM
This is kind of off the wall, but do you have to draw circles around problem areas? Would it be ok to have a can of spray paint, or something similar, to mark them? Seems like that would be easier than building an arm that can draw circles.

Joe
Title: Re: Need Help To Design Roofing Robot
Post by: Soeren on June 22, 2011, 03:35:34 PM
Hi,

1. It must be a mobile remote control device that is able to initiate small tasks while moving on sloped or flat surfaces and negotiate valleys.
Definitely a large tri- or quadrocopter with either wheels or whats-the-name (a helicopters "landing gear") with high friction rubber - wheels may be a better option in some instances... I think.
If wheeled, the propellors can be run  with enough power (and direction) to place and keep it where it needs to be. If needed, the propellors can even make it appear heavier, if you need that for digging out a shingle or whatever.


2. The main tool to consider is a quality camera (around 10 megapixels at least) that can preferably be seen from the ground on a laptop.
Only one I can think of... Canon makes a couple of DSLR's with WiFi, so you can frame the shots (and view them real time) on a remote (within reason) laptop.
Unfortunately, they ain't cheap.
Not sure how rain/water protected these models are, but a cover could be made very easy.


Other tools optional are a pitch gauge, and shingle gauge, and an arm that can use sidewalk chalk to circle damages.
Could you post photos of these gauges (with measures)?
As someone mentioned, spray paint might be easier - a cheap air brush and a CO2 bulb (as used on a soda syphon) with an electric valve for "air" would make a nice little marker "gun", but it would be more permanent than chalk of course.
If it has to be chalk, perhaps ground chalk could be blasted similarly to the paint or as a last resort, a piece of chalk could be held in a sort of flexible spring.


3. It must be able to stay on high slopes without trouble, so the wheel type is important. Ability to handle at least a 10 pitch roof is preferable.
That should be easy with a 'copter.


4. Flight is optional, or a mix of flight and ground mobility.
You said it.


5. Reasonable pricing is preferable.
That is the one parameter that is really really counterproductive! Sure, I know you haven't got unlimited funding, but anything short of the WiFi Canon DSLR will give you crappy shots that won't tell the whole story.
A 'copter isn't cheap to build either, as you need a pretty good lift to carry a DSLR, gauges, marker and the manipulators that's needed.
You're not constructing a toy where it's alright if it breaks down every now and then, but a pro tool that must work flawless at all times.
Unfortunately that costs a bit. Even if it was mass produced it would be expensive and you want to build a single unit - you sure wouldn't like the price if you asked an engineering firm to develop it.

Your best bet at keeping costs down, if you are up to it, will be to do all of the physical work yourself, obviously, but then you need to be at least fairly proficient in mechanics and be at least able to follow and understand an electronics schematic and finally program a microcontroller (or learn quickly).

Might sound like a hard journey, but the upside is that you get to select the materials - Carbon fiber, Kevlar, or Twaron fibers, titanium etc. is fairly reasonably priced if you just want raw stock.


6. Any way to view what the robot is doing from the ground is preferable.
Canon!
And that's from a guy who prefers Pentax (who unfortunately doesn't make remote framing cams) and for some marginal stuff Casio.


7. It needs to take measurements of the roof
Photos with a reference object is a fairly accurate method.


8. It should withstand hot summer heat and icy cold winters.
For how long at a time?


I hope you get your own personal job security 'bot and don't be afraid to take the leap, one user here (Tommy) has gone from zero to a substantial do-"all" giant in a very short span, by expanding bit by bit (and next time he writes, I'm sure it's a guard dog as well ;D).

You can get lots of help here, but ultimately, it's a question of how well you can construct it, so I'd like to ask you about your previous results with tools and materials.
Title: Re: Need Help To Design Roofing Robot
Post by: Roof Rover on January 09, 2013, 03:35:18 PM
We have built it.  Let me check the forum rules about what is ok to post.  Feel free to pm me.