Quadcoptor for 3D mapping of environments.

[Srijal]

Hello,

I'm a student of Electronics Engineering and I've taken up a project wherein a quadcopter is used to create a 3 Dimensional map of any Indoor space. The quadcopter will be manually flown around and will have carry Depth Camera (The Microsoft Kinect sensor) and an onboard computer to take images from the camera, process it, and send it to a base station using WiFi. (I'm unable to post links)

Now, I need help from you guys in deciding the parts for the quadcopter. I have scratch built an RC plane before and flown a few quads, and I have some basic knowledge of what it takes to build a quadcopter. But this application is very specific and I think someone with experience could direct me towards the exact configuration to go with.

It should be able to carry the Kinect sensor and Onboard computer (along with some peripherals for power, communication etc), which together do not exceed 1.2-1.5Kgs(2.6 to 3.3 lbs / 42 to 53 oz) of payload. High speed is not a requirement, it has to move and turn slowly for the map to form properly. Longer flight times would be great (at least 15mins). The budget would be around 150-250$ for just the quadcopter (not including the kinect and onboard computer).
Things I would need:

1. Frame, 400-600mm? (can build one myself to save cost, if necessary ?)
2. Motors and Propellors
3. ESC
4. Battery (the right mAh to weight ratio for maximum flight time)
5. Flight controller (Naze32/Flip32)

I'm going with the Naze32/Flip32 flight controller as it supports ROSflight, which is necessary as the entire system would be based on ROS (which provides a framework and tools for making such complex tasks relatively easy). Also, I already have the Turnigy 9x for the radio system.

Please point out any details that I've missed. Any suggestions would be great!

 

[Srijal]

So I did some searching and I think adding everything up, AUW would be about 3kgs (6.6lbs), which brings me to 3x2/4 = 1500grams of thrust per motor.

I've looked into some low kV (500-750kv) motors specs and few of them provide prop test data like the Hobbyking Donkey ST3508 -730kV which says:
12x6 - 11.1v - 16A - 1060g Thrust,
13X4 - 11.1v - 15A - 1140g Thrust,
10x5 - 14.8v - 16A - 1180g Thrust,
with the max current rating of 35A...so would it be able to handle a bigger prop (say 12x4) on 4S? But still, I don't think it will meet my requirements.

The main thing that confuses me is that what exact wattage/size am I looking for because even for a particular kV, there is an entire range of different sized motors and almost none of them have prop test data given (I'm using the Hobbyking motor selector tool). Any tips?

 

[yay]

none. i'm new and you're way out of my weight class. but i figure, if i post, then other people might be inclined to check the thread out.

 

[Srijal]

none. i'm new and you're way out of my weight class. but i figure, if i post, then other people might be inclined to check the thread out.

Haha, I appreciate your effort!

 

[Srijal]

I tried some configurations...the thrust to weight ratio is low but is it okay? What should I go about changing next?

 

[GJH105775]

I would now build knowing it would be that close. Try shedding weight or increasing thrust. Off the top of my head I think you can use a Turnigy Multistar 5200mAh battery if you were to get the 4S1P version. It would shed weight but has a lower C rating so be sure to verify it can support the current draw Then I'd look into making sure my prop and motor combination was efficient. I think 30 amp ESCs might do you just fine unless you change the drive to something that would be a draw more (might be a good idea to get more thrust).

 

[yay]

it seems that previous post is mistakenly tagged as a quote

i guess i do have a modicum of advice, from what i've *seen* of this pursuit, you ought to do quite well with a DIY frame. especially if the task is to hover around slowly. i've seen someone fly a couple of dowels taped together around for a half hour.. not quite what you want to moutn a pricey sensor on, but if you have any experience with making "structural" things (educated as an architect i think in terms of compressive/tensile forces) i'd think it would be a good place to optimise weight..

 

[GJH105775]

it seems that previous post is mistakenly tagged as a quote

i guess i do have a modicum of advice, from what i've *seen* of this pursuit, you ought to do quite well with a DIY frame. especially if the task is to hover around slowly. i've seen someone fly a couple of dowels taped together around for a half hour.. not quite what you want to moutn a pricey sensor on, but if you have any experience with making "structural" things (educated as an architect i think in terms of compressive/tensile forces) i'd think it would be a good place to optimise weight..

In the event of a crash you are going to break stuff. No matter how strong the frame you will probably break something even if it is not the frame. Generally the most common schools of thought in this are to either optimize for weight and controlling what will break, or preventing the accident as long and as much as possible then dealing with it when it happens. So shedding weight on the frame when possible is always great, as long as it does not increase chances of a crash. So dampen vibrations, use stiffer materials when possible to avoid flexing.

 

[Srijal]

I will be trying to shed weight as much as possible and I feel that 3kg (6.6lbs) is the worst case value for AUW. And for more thrust, I found this Turnigy L3020B-600 motor which seems to be able to provide a massive amount of power (600-800W). This brings me close to 2:1 thrust ratio but still the flight times are less when I'm using 4S1P config. I think I should increase the kV a bit with higher powered motors, that should do the trick...

@yay, I am comfortable with building a frame myself, what material do you suggest building it with? Aluminium? Plywood? Would that be better than the plastic moulded frames I could buy directly?

Also, see this video. That is the thing I want my quad to be able to do, what are the things you can judge by seeing the video?

Edit: Actually in the video, the mention the quadcopter is 'Ascending Technologies Pelican', which has its details mention on the website here. They don't have any specific parts named. What can be said from the numbers?

 

[Chuck]

none. i'm new and you're way out of my weight class. but i figure, if i post, then other people might be inclined to check the thread out.

I'm inclined to think that you need six rather than four motors for this payload, and the amount of time it will have to stay airborne. In order to keep from overloading the motors. Does anybody know if a Yuneec Typhoon H would be a good starting point for this project?

 

[Srijal]

I'm inclined to think that you need six rather than four motors for this payload, and the amount of time it will have to stay airborne. In order to keep from overloading the motors. Does anybody know if a Yuneec Typhoon H would be a good starting point for this project?

The cost would be a lot more, I don't think I can afford to spend that much. But considering lower power motors can be used instead of higher power ones for a quad, then it may be a possible solution...a hexacopter just needs 2 extra motors and ESC's right?

Also, I need to see if ROSflight can support that, which I highly doubt.

 

[yay]

i'm an "unconventional thinker" - it says so on my nametag and again, i'm real new at this.

today i scored about 4 meters of agave flowering stalks, which are probably the same volume-weight as balsa wood, but more durable. and sun hardened it's unlikely you can source such materials online as commercial materials are usually produced en masse, not single sourced from nature. but maybe you have some similar tough, light plant material available near you.

these stalks are quite rigid, being up to 4cm in diameter, and easy to carve. this is my best thinking on frame materials. also, when they break, you don't have to pay a lot of money.



didn't watch the whole video but that quad was oscillating a lot, from the considerable weight. some other approaches...

afaik (this is a big IF) you can run multiple motors on one ESC (???) so eg. an octocopter can have 8 ESCS and 8 motors, or 4 ESCs with two motors on each.

similarly, you might find better approaches having *dozens* of small motors.

and, if your scan region does not have to include the area directly above the sensor, maybe you are going about this entirely the wrong way i'm sure the project is supposed to involve the autonomous mobility as you have proposed, but .. if you can't just hold the thing in the air with a pole or suspend it from a cord, consider using a dirigible, or a combination of motors and eg. helium balloons???

only thinknig out loud. i get messy when i think out loud, so discard as suitable.

plus i know their pelican looks awesome with that big expensive snesor sitting on the top, but maybe wrappign a bunch of soft puffy stuff around the parts that don't scan would be smarter.

or jus twait a couple of years for someone to make a much more affordable, lighter sensor!

 

[Chuck]

The cost would be a lot more, I don't think I can afford to spend that much. But considering lower power motors can be used instead of higher power ones for a quad, then it may be a possible solution...a hexacopter just needs 2 extra motors and ESC's right?

Also, I need to see if ROSflight can support that, which I highly doubt.

Perhaps someone here with experience building drones with more than four motors, can answer that.

 

[Chuck]

i'm an "unconventional thinker" - it says so on my nametag and again, i'm real new at this.

today i scored about 4 meters of agave flowering stalks, which are probably the same volume-weight as balsa wood, but more durable. and sun hardened it's unlikely you can source such materials online as commercial materials are usually produced en masse, not single sourced from nature. but maybe you have some similar tough, light plant material available near you.

these stalks are quite rigid, being up to 4cm in diameter, and easy to carve. this is my best thinking on frame materials. also, when they break, you don't have to pay a lot of money.



didn't watch the whole video but that quad was oscillating a lot, from the considerable weight. some other approaches...

afaik (this is a big IF) you can run multiple motors on one ESC (???) so eg. an octocopter can have 8 ESCS and 8 motors, or 4 ESCs with two motors on each.

similarly, you might find better approaches having *dozens* of small motors.

and, if your scan region does not have to include the area directly above the sensor, maybe you are going about this entirely the wrong way i'm sure the project is supposed to involve the autonomous mobility as you have proposed, but .. if you can't just hold the thing in the air with a pole or suspend it from a cord, consider using a dirigible, or a combination of motors and eg. helium balloons???

only thinknig out loud. i get messy when i think out loud, so discard as suitable.

plus i know their pelican looks awesome with that big expensive snesor sitting on the top, but maybe wrappign a bunch of soft puffy stuff around the parts that don't scan would be smarter.

or jus twait a couple of years for someone to make a much more affordable, lighter sensor!

"We all live in a yellow submarine, a yellow submarine, yellow submarine...."