+-5mm quadcopter to ground system

Hi all,

I am trying to make a quadcopter for construction surveying staking for my business. I have figured out all the different parts I need to build it, but one thing remains. I need a very accurate system to get distance from a wooden hub in the ground (accuracy +- 5mm) to my quad. Accuracy is VERY important for this application. Does anyone have ideas for what I can use. If I left out some information please let me know and I can give it to you. I am new-ish to arduino but have lots of programmer/EE friends so difficulty of code isn't necessarily a factor.

Thank you

You don't mention over what distance you need it to operate, but I guess a few hundred meters? What's your budget? Its' not going to be cheap.

http://hackaday.com/2013/08/05/centimeter-level-precision-gps-for-500/ http://www.kickstarter.com/projects/swiftnav/piksi-the-rtk-gps-receiver

Shpaget,
Thank you for the quick reply!

My range would be <10 feet. That would be from the bottom of the quad to the wooden hub.
I don’t know what a system like this would go for, but I have business money so anything reasonable (less then $1000?) would be acceptable.

I have seen the piski GPS and actually was already planning on buying that, but I didn’t know if it could get me accurate altitude of the quad. With normal surveying equipment you use a rod that is a fixed distance in height so curvature of the earth is accounted for.

Ok, I had an entirely different picture of your requirements. I thought you'd have the quad fly around the construction site and measure... stuff. If you need it only to hover above your hub at 10 ft then the stick would be a better alternative. Why not use it? Getting a +/- 5mm precision is hard enough, station keeping within those limits is even harder. Wind can blow the copter away quite a bit.

So I am obviously missing something. What is it that the stick can't do that you need to copter for? Can some equipment be mounted on the stick to address those shortcomings?

The efficiency of the project would be raised considerably. With the current way of doing things you would have 2 people operating equipment that costs $40K and take on average 10-15 min to set 1 point. My cost for the system right now is $3K and only paying one person to do the job, giving me more crew to stake more sites.

I realized I did not specify the goal of the project, I apologize for that. Here are the goals/operations I am needing: 1) quad able to fly to a set coordinate point (this can be done with the Piksi) 2) quad able to determine the distance to the ground. (I will have a BMP085 to give me an estimate of altitude for flying to points but for the quad I need a specific sensor, LIDAR, IR to give me distance to the ground to within 5mm give or take) 3) after taking the quad to ground reading I need to be able to compute a cut/fill of that individual point to a set construction elevation (I have already accounted for this too) 4) quad able to fly to the next point, rinse, repeat

-Michael

It might be possible using ultrasonics, by measuring both the time-of-flight and the phase shift. I suggest you work out what effect the wind would have on the reading before going down this route.

You may be able to find a laser rangefinder module with that accuracy, but it will be expensive and probably too heavy.

+/- 5mm accuracy over 120 inches is like +/- 0 .1% accuracy/resolution. All from a moving/vibrating airborne platform. Not saying it can't be done but I'm of the opinion that if you can't find some commercial firm that has already accomplished and has a product available at whatever costs, your chances are pretty slim in my opinion. Laser time of flight with lots of averaging/filtering is probably the only sensor that might have the repeatable accuracy needed IMO and it would need a good reflecting surface on the target to bounce off of. But again I don't think you would find the quadcopter to be stable enough to obtain that level of accuracy.

Lefty

dc42, interesting idea I will look into the feasibility of the ultrasonic system and wind impact on readings. That is an option I have read about but didn't do much research into.

Lefty, you bring up good points about the demands that the laser would be under. About this laser TOF system, is this the same type of system that is in laser measuring devices that construction companies use to gauge distances for cutting wood? something like this? http://www.zorotools.com/g/00052806/k-G1580887?utm_source=google_shopping&utm_medium=cpc&utm_campaign=Google_Shopping_Feed&kw=%7Bkeyword%7D&gclid=CM75x4Tct7sCFU_NOgod3T8Aag

this has the accuracy that i am needing, so I'm sure the parts are out there, i just don't know how to find them.

-Mike

mjb3063: The efficiency of the project would be raised considerably. With the current way of doing things you would have 2 people operating equipment that costs $40K and take on average 10-15 min to set 1 point. My cost for the system right now is $3K and only paying one person to do the job, giving me more crew to stake more sites.

The problem is, you're asking for military precision, and you're not gonna get that on an aircraft of any kind, for less than that $50K you're trying to beat. I don't even think you can get that with $50K - trust me, I've worked in aerospace with this kind of stuff, and while there are COTS LPS systems you could use, they aren't below your price point. This is not something drones are going to be good for. Drones are good at covering large areas quickly - they are not good at precision stuff.

Furthermore, a multi-copter of any kind is never going to be stable enough for surveying in the traditional sense. A multi-copter might be useful for laser surveying where you don't need lots of accuracy but need to get the general lay of a large area. Trust me on that - I fly quad copters and hexcopters every day. I have a friend who does aerial photos as sort of a business, and guess what his favorite aircraft is? 30-foot pole.

It might be possible to get a quadrotor solution working, but I recommend you implement a hand held version first. (IMO the quadrotor part of the solution adds an order of magnitude more complexity and some health and safety risks that strike me as significant.)

As far as I understand it there are already solutions for this sort of thing based on an instrument pod on a stick that beeps to tell you when it is at the right level. Something similar to give you feedback in three dimensions seems easy enough - I'm envisaging a telescopic pole that you would carry around until it told you that it was in the right place, then rest the bottom on the ground and extend the pole until the head was at your target height - the amount of extension tells you how high or low the ground is at that point.

You should look into photogrammetry for this. Here’s one example, which claims accuracy of .0312":
http://www.etemplatesystem.com/Default.aspx?tabid=1154

Thank you all for the advice. I have found companies that offer COTS laser TOF systems. Seems that +-5mm is a little stringent but they think it will work if I use a highly reflective surface and low altitude.

mjb3063: Thank you all for the advice. I have found companies that offer COTS laser TOF systems. Seems that +-5mm is a little stringent but they think it will work if I use a highly reflective surface and low altitude.

If you mounted it on a stabilized gimbal that might work. The problem we've been having with using similar systems for altitude stabilization and auto-landing is, the angle of the copter can change the sensed altitude dramatically. Even when hovering over a perfectly flat surface, two degrees off-angle can have you suddenly thinking you're 10 feet higher. So the difficulty has essentially been 'noise' on the sensor, caused by the fact that the quad copter is an unstable platform. If you could remove the instability of the copter, you could get a less noisy distance reading. You could average distances over time, but that's not going to be the precision you need.

It's a very interesting idea for sure - I like the photogrammetry idea personally. If the sensing time for that is very short, then instability in the platform might not matter. Standard photographs from quad copters are usually fairly sharp, so if it's a regular exposure time, it should be fine.