line folowing aerial vehicle help needed

[wardhanster]

well as you all have probably predicted what this thread is all about...
i want to Make a line following aerial vehicle which can follow a line on the ground. Important thing is the areal vehicle is to perform in a closed room or may be an auditorium,the limitation of the flight height is abt 2.5m... any ways what i have in mind is to use the concept of a line following robot and use it to follow a line but the questions which are making things difficult are ..
1> the platform i shud use, i.e whether the vehicle be an airplane or a hovercraft.. or a vtol aircraft.
2> what sensors i have to use in order to precisely detect the line on the ground.. the vehicle will be following the line at a height of 50 cm.. 2.5 m is the upper limit

my background details are ...
involved in the hobby of r/c plane flying and construction for past 6 years.. so designing a airplane is not an issue..on the other hand just finnished my 1st project of the line following robot and hence have enough knowledge for setting up and controlling things with a microcontroller...
help me guys

[MangoBot]

for the platform make it a helicopter, because it would be able to go slowly and hover, Im not sure if a IR pair would have that range. Lets hope smash replies

[Finnik]

I personally don't think an IR emitter/detector pair would have the accuracy to follow a line at 50 cm altitude. Of course, we are missing some vital points of information: the width of the line, the colour of the line and the colour of the surface the line is on.
You could also look into using a small camera.

[wardhanster]

these are the details regarding the path to be followed by the vehicle
___
white line on a dark background
The path could be in any reasonably curvy shape ( without sharp turns and without criss-cross ).
There would be a white line of 10cm thickness (on a black background).
The arena is a dimly lit hall .
including an camera will be out of the budget and may be out of my programing skills .. told ya guys i have just started.. these

[SmAsH]

Arg, this is going to be one hell of a ride to complete!
personally i would say a helicopter like base (tricopter, quadcopter etc...) would be the easiest to keep upright
and steady, with a plane, i think it would go way to fast for this...

For sensors, i don't really think an led will have that kind of range at all, it will struggle to do 30cm decently!
if you have the money (you have to for this project!) you can look into stuff like blackfin, cmucam and avrcam.
if money is kinda tight and your stretching it, an ir rangefinder may suffice.

but i think a project like this would need to be divided into steps, first you will need to look at what others have
done, their control systems and the shapes of their aircraft.
the next step would be designing etc...

Lets hope smash replies

HAH! why? i have no personal experience in r/c planes/helis at all, never even flown a plane!
i only know what ive read from others.

[Razor Concepts]

I think the major problem here is the flying vehicle part. Airplanes are out of the question as they are not indoor types (you can get smallish ones but no way to put electronics on those). Helis will take quite a bit of programming for stable flight, kalman filtering, gyros, etc. Maybe get a coax like cx2 or big lama for their natural stability properties.

[sonictj]

If your gonna make a flying robot I think it makes more sense to follow gps way points.  You can do that with a plane outside.  My advice is that you need to focus on the autopilot and then whatever task you want the robot to perform.  making a robot fly itself is a very daunting task.  Mostly because bad programming = crash and likely a broken robot.  Comparatively ground robots are generally robust and do not require such care in programming, design etc.  Don't let that stop you, just think this through.

[docel]

Looks like you're preparing for the Shaastra Aerobotics competition.....

I personally feel that the Problem statement is a little too superfluous for a student project. As usual, the competition will end up with no true contenders and the rules will be changed impromptu.

Let us examine the problem: Since no dimensions are posted, lets assume the shaastra event as the base:-

1. Line following is difficult for a flying object- the more it is above ground the more unreliable it will be.
2. Normal Line following techniques with conventional sensor systems become incompetent , as the distance up from ground increases.
3. The machine must have a very fast reaction time indeed.

Since the dimensions have not been mentioned, we must assume that the Flying object should be as small as possible.

 If it is small, then it should be quite powerful to take the payloads of sensors and the additional electronics for line following/ flight control.
 As the dimension of the line is not known, one must consider the possibility  of a " Line "  < 10% of the bot size, or a " Path "  which is >25% of the bot width.

In general, Flying objects - whether Plane, Heli or G effect machines are highly unstable in nature. So, the maximum effort is towards the construction and control of the machine, before sensor management techniques.

To summarise::

You need a Bot that must do these things-

- Small , Light , Powerful and VERY stable.
- capable of slow flight
- fast sense and control design

This calls for a Vertical flight system rather than a Horizontal flight system as far as lift is concerned.

A Hovercraft fits the bill quite nicely indeed and is the only easy way our for an amateur/ student implementation.

Reasons??

1. It has natural Height control and will maintain it quite accurately.
2. Similar to a Ground vehicle so the wheeled line following techniques are enough.
3. Weight and size does not matter at all
4. The problem is the Skirt- but it may not matter....


Good Luck !!!

[mdmedlin]

This calls for a Vertical flight system rather than a Horizontal flight system as far as lift is concerned.

I agree in the fact that the hovercraft has the most stabel base and will fly the slowest.  I am not sure if the hovercraft will make it 50cm off the ground though because of the skirt issue.  If you do end up making a hovercraft, it will basically be a line following ground robot but sans the wheels.  All the programming and materials will be the same except that instead of moving wheels, you will have to change the direction of the air blast.  I hope this helps

[wardhanster]

Looks like you're preparing for the Shaastra Aerobotics competition.....

I personally feel that the Problem statement is a little too superfluous for a student project. As usual, the competition will end up with no true contenders and the rules will be changed impromptu.

Let us examine the problem: Since no dimensions are posted, lets assume the shaastra event as the base:-

1. Line following is difficult for a flying object- the more it is above ground the more unreliable it will be.
2. Normal Line following techniques with conventional sensor systems become incompetent , as the distance up from ground increases.
3. The machine must have a very fast reaction time indeed.

Since the dimensions have not been mentioned, we must assume that the Flying object should be as small as possible.

 If it is small, then it should be quite powerful to take the payloads of sensors and the additional electronics for line following/ flight control.
 As the dimension of the line is not known, one must consider the possibility  of a " Line "  < 10% of the bot size, or a " Path "  which is >25% of the bot width.

In general, Flying objects - whether Plane, Heli or G effect machines are highly unstable in nature. So, the maximum effort is towards the construction and control of the machine, before sensor management techniques.

To summarise::

You need a Bot that must do these things-

- Small , Light , Powerful and VERY stable.
- capable of slow flight
- fast sense and control design

This calls for a Vertical flight system rather than a Horizontal flight system as far as lift is concerned.

A Hovercraft fits the bill quite nicely indeed and is the only easy way our for an amateur/ student implementation.

Reasons??

1. It has natural Height control and will maintain it quite accurately.
2. Similar to a Ground vehicle so the wheeled line following techniques are enough.
3. Weight and size does not matter at all
4. The problem is the Skirt- but it may not matter....


Good Luck !!!

well you got me there, yes indeed i m looking forward for this event shashtra ... but personaly this problem statement just got personal... lol..
any way... what i was thinking a height of 30cm or less (anything that does not touch ground ) will be enough for detecting a line of 10cm in width on a dark background... my idea is to detect the colour of the line and the background... this gives the room for fuzzy logic ... what i mean is that i will be using 4sensors, two sensors will be at 4cm from center of the vehicle... this means 8cm from each other(for detecting the line colour)... and other 2 sensors at 11cm(for detecting the background colour) from each other..... and the old and good line following technique....
this is not the final idea .. but i think it will be a good base to start with...what i was wondering it the sensor that i shud be using.. is that the traditional ir and photodetector pair will do(that line follower usess) or some thing else...

[Ro-Bot-X]

CMUcam has connectors for pan/tilt servos (as far as I can remember) so that could be used for actually control the aircraft (perhaps little mod of the firmware is needed) or just connect it to a Ardupilot. The camera has to be mounted high on the wheeled robot, so it will go nicely on the aircraft. Does a Blimp qualify?

[wil.hamilton]

i was also thinking a blimp would work best.  with a blimp you don't have to stabilize it (it does that already) and the control systems are more akin to skid steering.  because you're detecting the line at such a long range.  IR sensors most likely wont work.  your best bet would be to use a cam of some sort, like smash said, the blackfin, cmucam, and avrcam are all good choices.

[docel]

  A Blimp fits the bill quite simply, indeed.....but, what the OP did not mention is that there is a size ( 40cmX40cmX40cm) restriction that eliminates even winged aircraft ( which needs a higher Payload capacity if it must carry sensors, Cameras etc.,).

There is not much choice left here.

 Sensors will depend on the height from ground. The rules does not specify the "minimum" height, unless they change the rules overnight ( as they usually do at Shaastra ) so anything from say 1cm to the maximum height must be decided first.
 Visible or IR sensors ( an array!! ) can be used, if one is careful with the field of vision. Some optics will be necessary to ensure this and can be done. We got our 'Flier' up to 2ft. with an 8 LED visible light sensor array at at a slow forward speed of 10cms/s. But its quite difficult and painful, though.

[WhomBom]

you could also use an single line camera, it has a resolution of for instance 128X1 pixels and can detect shades of grey Focussing point could be adjusted, i just can't remeber where i read it

[hgordon]

Use a blimp.  Here's an example of a system that has the capabilities you need -
   http://www.surveyor.com/YARB.html

Start with the Blimp Guys 66" envelope - it will have enough lift for whatever electronics you choose -
   http://www.rcguys.com/66blimp.html

Here is a smaller blimp that may or may not have enough lift, but it is cheap -
    http://diydrones.com/profiles/blog/show?id=705844%3ABlogPost%3A44817