Building a Flying Robot

[airman00]

no i mean a two bladed chopper like this
top view would look like this
O-O

So it would create a Chinook

exactly like a chinook!

any robot chinooks out there?

@Kawes  : thanks for the video!  Any idea  how to make it not spin the body so much , or maybe you can do three or four helicopters taped together , please?

[KaweS]

believe me that i'd love to, but i;m just not allowed to do sth like that ;p
my mgr was just off today. i didn't expect that so i hadn't too much time to figure out a nice connector for them.
next time i'll try to give you some more ;p

[hgordon]

Next time, use coaxial rotor heli's like the Blade CX2 and you won't have the issue of rotational torques.

[KaweS]

well... we do just one rc heli with coaxial rotor - Apache AH64 Longbow
the only problem is that i am not allowed to take it out of the box, charge it and play ;[
the same with those Picoo Z - they came from returned packs, where the other one was broken.
i use what i can, but i'd love to be able to 'test' some better stuff ;p

[airman00]

well... we do just one rc heli with coaxial rotor - Apache AH64 Longbow
the only problem is that i am not allowed to take it out of the box, charge it and play ;[
the same with those Picoo Z - they came from returned packs, where the other one was broken.
i use what i can, but i'd love to be able to 'test' some better stuff ;p

OK cool!

Im thinking of having four helicopters attached together ( with the back rotors taken off) , maybe you can do that one day ......

[KaweS]

i just read that IR may have problems to recognize black colour - id doesn't see it!
does it always occur?
does it depend on quality of IR and it's readings?
can i adjust it somehow?
what about working in places with various brightness? let's say it flies from building outside.

i think i gonna need to use both IR and sonar to determine space availability...

and what about reading through rotors. is it possible with IR? do i need to place it on a boom or sth?

[KaweS]

last night a CAD a big my robot
i'll write something after work
just 2 pics for now:

                 overview                                                  radar [on top and bottom]

what ya think about such a design? i forgot to attach legs [3 or 4]

PS: propellers are bent wrong direction - but it has to be just a general design to give you an idea of what i'm thinking of ;p

[Tsukubadaisei]

I am building some flying robots as well. http://www.societyofrobots.com/robotforum/index.php?topic=3197.0

But no offence but from your CAD I dont believe your robot is going to fly. I think the body in the middle will reduce a lot the efficiency of the two motors. Have you tried to calculate the flux of air through the propellers. The fluid lines(I dont know how to say those in English)? Not to mention that that body is difficult to build. I am not a mechanical engineer so I cannot expleing everything in detail to you but as a bio-system engineering student I studied some basic mechanics at college including fluid-dynamics. And I can tell you that in order to build high-speed, flying or under-water vehicles and robots, some basic knowledge of flui-dynamics is fundamental. I am not telling you to buy the thickest book in the book store about fluids. Just use google and get some idea about aerodynamic principles. I think ADMIN knows about this stuff much better than me. He is a mechanical enginer and is building an aquatic robot.

[Trumpkin]

you need a waaaaaaay bigger propeller on top to keep it up. also i don't think the propeller on the bottom is a good idea. i think if the propeller on bottom is even a little bit slower (or faster) than the one on top it will do flips. Also the body is to fat and will catch wind to easily.

[gmatkins]

It looked to me like the propeller on the bottom is intended to counter-rotate to alleviate the need for a stabilizing rotor. And depending on the resources available the body is doable. I think this thing will fly.

[Trumpkin]

it may fly I don't think it will. also, wouldn't it be confusing when the robot flys since you wouldn't know which is forwards?

[HDL_CinC_Dragon]

Its an intriguing design indeed. I think if you make the diameter of the propellers bigger, maybe even double, and made the body a little bit skinnier you would have more chance of a successful build. You may have difficulty turning and such.

[Trumpkin]

I think if you make the diameter of the propellers bigger, maybe even double, and made the body a little bit skinnier you would have more chance of a successful build. You may have difficulty turning and such.

that's pretty much what I said.

[HDL_CinC_Dragon]

I was just elaborating and building upon what you said...... and im cooler *Strikes a pose*

[KaweS]

cheers for those replies
to be honest i have no clue if it would fly - this is just an idea how it would look like. i've got another idea, but i'm too tired to CAD it today - maybe tomorrow.
for sure i don't want it to look like a helicopter - UFO-like thingy is what i want ;p

I am building some flying robots as well.

hope you'll help me ;ppp

But no offence but from your CAD I dont believe your robot is going to fly. I think the body in the middle will reduce a lot the efficiency of the two motors. Have you tried to calculate the flux of air through the propellers. The fluid lines(I dont know how to say those in English)? Not to mention that that body is difficult to build. I am not a mechanical engineer so I cannot expleing everything in detail to you but as a bio-system engineering student I studied some basic mechanics at college including fluid-dynamics. And I can tell you that in order to build high-speed, flying or under-water vehicles and robots, some basic knowledge of flui-dynamics is fundamental. I am not telling you to buy the thickest book in the book store about fluids. Just use google and get some idea about aerodynamic principles. I think ADMIN knows about this stuff much better than me. He is a mechanical enginer and is building an aquatic robot.

i don't know anything [yet] about aerodynamics and sure i'll google it first. if not - i'll have thousands of experiments for a good body ;p

also i don't think the propeller on the bottom is a good idea. i think if the propeller on bottom is even a little bit slower (or faster) than the one on top it will do flips.

i think if one is slower than another, then it will turn, and that's what i want to use to operate it

Also the body is to fat and will catch wind to easily.

for sure i don't want it too big. i'm still not sure if i want to fly it outdoor. if so. it needs to be quite heavy to force the wind. if not - i want it as small and light as possible.

It looked to me like the propeller on the bottom is intended to counter-rotate to alleviate the need for a stabilizing rotor.

exactly - it's a kind of coaxial rotor ;p

And depending on the resources available the body is doable.

i think to make body of polystyrene - easy to adjust, with some kind of cage to make it a bit stronger - plastic perhaps.

it may fly I don't think it will. also, wouldn't it be confusing when the robot flys since you wouldn't know which is forwards?

i'll paint eyes and mouth on it ;ppp hehehehe
well.... what's the difference for me? it's important for robot to know where is it's front ;p


i'm just surprised that no one said a word about the radar... i read a bit about them and i know that they're better than sonars in the air. but i couldn't find anything how could they work in small areas and at short distances...

for now i just keep on learning anything what is helpful to build a flying robot ;]

cheers,
Karol

[skatj]

Even if it does fly, that thing will be as stable as a jet-propelled brick, which would require hundreds and hundreds of dollars and the most difficult coding imaginable to keep it in the same place for more than half a second.

With two rotors, top and bottom, how do you intend to control the direction of flight? I'm not an expert on coaxial rotors, but I believe yaw can be controlled through differential torque of the rotors (doubt it would work on something like that though), but that won't matter unless you can pitch or roll the craft as well (can't tell which one it is for that design).

Don't try to come up with your own innovative design, especially if you don't know basic aerodynamics. Copy what others have done, you can't expect to pioneer anything with no experience.

A gps-only fixed wing aircraft designed around an inherently stable airframe, with a manual override pilot to take off and land is the cheapest, and easiest aerial robot to build.

Oh and, you won't be able to use a cheap accelerometer for something like this. You need a full-blown 6 axis IMU, especially since you want to design a completely new airframe that is incredibly unstable.

[hgordon]

Even if it does fly, that thing will be as stable as a jet-propelled brick, which would require hundreds and hundreds of dollars and the most difficult coding imaginable to keep it in the same place for more than half a second.

+1 what Skatj is saying

If you're serious about this, spend some time reading the threads on http://www.rcgroups.com/uav-unmanned-aerial-vehicles-238/ and http://www.rcgroups.com/vtols-360/

[KaweS]

it was just a proposition - i don't know how will it look like at the end, but certainly not like the first design - it's ugly ;ppp
second one is on it's way

how do i want to move?
first of all i don;t know yet if i gonna make variable pitch of blades - it will be more difficult and at the moment i don't know if it's worth effort
1. hovering
both rotors spin at the speed with opposite directions
2. moving up/down
adjusting pitch of blades
OR
adjusting speed of rotors
3. turning
one rotor spins faster and the other spins slower
4. moving forward/backward
and here i have problem ;p
general idea is to change pitch of one or both rotors, but i have no clue how it would behave - if i won't find anything i'll just test different solutions
what i'm thinking about is to tilt top rotor or both rotors that direction [it depends how fast the robot would turn] and then tilt the rotors to normal state.
i was thinking about a servo for each of rotors, to tilt them. you should already know - it can NOT change direction while is moving. but it's not problem at all - there's an other problem - keeping it balanced would be impossible or very difficult to achieve.
so i think i need to design a swashplate. it's difficult as well, but far less than previous solution... ok it was not a solution ;p

btw - what do you think about 2 motors connected parallel? does it have a sense?
 
[click to enlarge]

[skatj]

Again, what are you prepared to spend on this project? Even tried-and-true rotorcraft airframes that are relatively stable cost hundreds of dollars and extremely complex coding to balance. If you are deadset on going with rotors, at the very least go with a quadrotor or coaxial helicopter design, but I warn you even the most skilled of amateurs (read: non-military, non-industry) have not completely succeeded in building an autonomous helicopter. No offense, but what makes you think you can do it?

Variable pitch rotors are EXTREMELY difficult to fly, even for human pilots, let alone an autopilot.

I am telling you this because I too, once wanted to build a UAV, and I was just as naive and uninformed about it as you are. I had people who gave me a dash of cold water and put me on the right track, and if they didn't, then I would have ended up wasting hundreds of dollars.

[KaweS]

i am ready to spend £££ or even ££££ in years to design and then develop it. i am ready to spend thousands of hours for research.

what about complexity of coding? i really don't understand - how could it be so difficult? i guess what all i need is just to make it able to go any direction in 3D and if it drifts anywhere - go back.
ok - maybe i wrote it too easy. i know that i need to include present state of the rotors and even every single blade in it. i know that i need to stop before i achieve the destination, to stop exactly there, where i want to. and so on.

why did i choose a design with a rotor on each side? cause i found it the best solution to keep it small - coaxial needs a lot space between rotor discs - otherwise they will collide with each other. i know that would be MUCH easier - i could even use a ready mechanism from an rc heli. but it's not what i'm looking for.

and what make me think that i can? simplifying i think that i can make it, cause i think it's just up to me [and money] ;p it's extremely difficult and that makes me even more into it. as you know i'm not experienced. i am hungry to learn more and more about it. i believe that there is no impossible...
if i'll fail... well bad luck ;p but if i'll succeed... i'll be in kinda heaven ;]

as i guess you have a bit of knowledge. can you then help me? at least point me, where to look for or even what to look for.
what about money? i'm not a typical human - i don't like money... i'm dreaming about peace in the world and as long as there will be money there will not be peace. even if i'd put ££££ in it and then fail i wouldn't care about the money, but about failure itself.

[HDL_CinC_Dragon]

For the "if it drifts anywhere, go back" part you would need to use GPS and have the robot monitor its GPS coordinates and if it sees a change in its latitude, go the opposite corresponding direction, etc, etc. This isnt quite as easy as it sounds. Programming can be quite difficult for that sort of thing. Im too tired to get into detail as Ive been up for too long on too little sleep >_<

I think the easiest way to go would be a standard chopper design. That way, up and down is controlled only by one rotor, and turning is controlled by another. Both independent of each other..... tired. must stop typing.

[KaweS]

For the "if it drifts anywhere, go back" part you would need to use GPS and have the robot monitor its GPS coordinates and if it sees a change in its latitude, go the opposite corresponding direction, etc, etc.

as far as i know hi-end GPS receivers can pinpoint object to 2-5 meters [i need 1cm or so]
i though gyro/accelerometer would solve the problem [correct me please if i'm wrong]

I think the easiest way to go would be a standard chopper design. That way, up and down is controlled only by one rotor, and turning is controlled by another. Both independent of each other.....

for sure, but as i already told - that's not what i'm looking for

tired. must stop typing.

me too... g00d night and looking forward for few more words when you wake up ;]

[izua]

isn't a bee a pretty good flying robot? has tons of mAh-s, enormous torque and speed (think wings flappings per second) and is pretty smart just go get a bee and put it in a box

[airman00]

OK guys, this thread is starting to get sorta off topic

the first thing  is first

how will the darn thing hover autonomously?

Once we get that set straight , then we'll talk about GPS and sonar ,etc.

I propose three blades , but maybe there is a better way

After doing some research on the chinook helicopter I found that it works like this

The Chinook is powered by two turboshaft engines, mounted on either side of the helicopter's rear end and connected to the rotors by driveshafts. The counter-rotating rotors eliminate the need for an anti-torque vertical rotor, allowing all power to be used for lift and thrust. If one engine fails, the remaining engine can drive both rotors.[2]

It has two blades on top for lift and two jet engines pointing backwards for thrust forward

Maybe we should make two blades for thrust and two fans in back for thrust...what do you think?

[KaweS]

OK guys, this thread is starting to get sorta off topic

god bless you ;p

how will the darn thing hover autonomously?
Once we get that set straight , then we'll talk about GPS and sonar ,etc.

i wrote that before. for now we can skip starting the robot - i'll launch it from my hands at the beginning, till hovering and moving will be done.
once it's in the air, it'll make decisions based on sonars reading - if it drifts any direction, it will try to go back by calculating proper angel of rotor discs and the thrust. while it's moving [or trying] ;p it will still keep on calculate readings and include the previous ones. on that base it will decide to adjust angel and thrust.
that's why i ask for sensors - which will be the best for this purpose? sonar/radar/IR or any combination of them. i'll repeat once again, that GPS can pinpoint to metres - not centimetres, what makes it useless for me. if you know any better GPS devices - let me know please ;]

I propose three blades , but maybe there is a better way
After doing some research on the chinook helicopter I found that it works like this

The Chinook is powered by two turboshaft engines, mounted on either side of the helicopter's rear end and connected to the rotors by driveshafts. The counter-rotating rotors eliminate the need for an anti-torque vertical rotor, allowing all power to be used for lift and thrust. If one engine fails, the remaining engine can drive both rotors.[2]

It has two blades on top for lift and two jet engines pointing backwards for thrust forward
Maybe we should make two blades for thrust and two fans in back for thrust...what do you think?

This configuration also has the advantage of being able to hold more weight with shorter blades, since there are two sets. Also, all of the power from the engines can be used for lift, whereas a single rotor helicopter uses power to counter the torque. Because of this, tandem choppers are among some of the most powerful and fastest. The CH-47 Chinook for example, has one of the fastest top speeds of any helicopter in service.

the third is my favourite sentence ;]
but it's too big for me...
general idea of using 2 horizontal rotor discs to get as much thrust as i can get from 2 engines is still included. also i included idea of coaxial rotors. but it's still too big - they can NOT be to close to each other, because they will collide with each other when you'll tilt one [in real helicopters just the top rotor change pitch]
i want to keep it really small and that's why i've chosen such design. and i don't see a reason why it would be [far] less stable than normal coaxial helicopter.

so please - if you can help me - help me
stop saying that i will not do it cause it's too difficult...

cheers,
Karol

[airman00]

it has to hover in place


that is a must , you can't have it just flying around everywhere. Hovering is a must!!! that should be done first!!!!

what about this design?


You can have a servo tilt the blades

[KaweS]

it has to hover in place
that is a must , you can't have it just flying around everywhere. Hovering is a must!!! that should be done first!!!!

eee yeah i know. i didn't say i want to skip it. i meant i can skip start from the ground at the beginning till it will learn how to hover and then i'll go back to take off.

what about this design?

transverse rotors are like tandem - they take too much space ;[

You can have a servo tilt the blades

i thought about it as well, but in this case i'd need 2 servos for each rotor disc to tilt it any direction. it's much simpler than a swashplate, but drain more energy and much more room. still, i didn't refuse it.

[Trikky2]

Hi there,

If you are interested in buying a kit of parts to start a fling robot how about something like :
http://www.asctec.de/main/index.php?id=15&pid=14&lang=en&cat=hobby

The people across at surveyor.com got one flying with the 'guts' from their SRV-1 controlling it and the camera working.

As for using the coanda effect, have a look at :
http://jlnlabs.online.fr/gfsuav/index.htm

All the best.

Richard

[KaweS]

i've already seen it, but thanks anyway
i didn;t know about coanda effect and i need to learn a bit about it as well ;p

cheers,
KaweS

[paulstreats]

I remember a demonstration of this effect is to place a small paper bowl upside and quickly waft your hand over the top of it. The bowl should lift up.

A paper bowl with a computer fan over the top of it might be a more promising demonstration, its all about passing air over the top of a curved surface. It creates a moment of low pressure, meaning that everything surrounding the low pressure area tries to fill it, so air from above below, left, right etc tries to fill the gap. In realism air from all the angles rushes into the area, but there is a drag towards the center of the low pressure. If there is an object there, this also gets dragged towards the area of low pressure, so the fan creates a low pressure area and the rest of the craft gets dragged towards that area (remember that every other direction gets dragged also). I pretty sure that this is how it works, but ive not heard mention of it for 4 years or so.

It might be interesting to see ideas using impellor type fans aswell