Hi
I need some input for solving a "problem" with flying a multirotor in a tube (dont ask why ;-) )
When flying manually in a tube/pipe with relatively short distance from propellars til the tube-wall there seems to be a vacuum spot in the center of the tube. The vehicle is being pushed up, down or to the sides.
I am considering if an arducopter setup with lidar sensors would be able to get the preferred stability and entered flight?
Any thoughts?
/peter
Replies
Hi , I am working on a same problem, since the flight in a pipe cannot be controlled thru radio signals I want to make an autonomous navigation, also using a lidar
https://www.youtube.com/watch?v=d-u0w0hC9s8
How about tilting the motors outward (5-10 degrees) - a kind of dihedral effect - to direct the prop wash away from the craft (somewhat forward and backward along the pipe)?
I believe it helps solve the bouncing effect when trying to land, by reducing the pressure buildup below the craft (and resulting interference with barometer).
At the cost of some loss of flight time/lift.
Or, how about something like this:
and just go with the flow. Bounce around a bit. (because sooner or later, your 'naked drone' is going to hit a wall, break a prop, and strand itself inside the pipe).
George
Hi George.
The platform should avoid touching the sides of the pipe.
We are actually testing the skewed motor tilts in the coming days. However I would still like to find out what proximity sensors would be best, and the FC.
Thanks, Peter
if you hover inside a narrow tube, you make much of the air raise on the sides, and go down in the middle(not really a vacuum) , you need climb rate that exceeds the top speed of the air (depending on the friction) - other than that, using any kind of sensors , combined with a fast control loop , should be able to fly under these conditions.
Thank you for the answer.
Do you know if it is possible to use 3 lidar sensors on for example the pixhawk? Or is there better fc's for this?
/Peter
lidars should work fine, of course, depends on the surface actually reflecting well enough. you would also like to shoot only one lidar at a time, to avoid possibility of reading reflections from each others light.
As for FC, you would absolutely need something that is open source (not "partially" open source like mikrokopter where navigation code is treated as a secret (as if it actually was good in any way)).
Then you have ArduPilot , and OpenPilot - any of those fulfill all requirements just fine. both will need some programming to do such navigation.
Thanks, Andre
In your opinion, are there better sensors than Lidar for close proximity use? The distance in this case from sensor to surface will be ~15-20 cm
thx
/peter
I do not know about any more suitable sensor than that.
Do you know if it will be possible to use for example 3 liar sensors to the pixhawk or similar?
The reason I am asking is because it is not clear to me if this is possible? And it is quite expensive to purchase 3 of them, especially if they can't be used ;-)
Thx
/peter
if you mean laser range finders, connected to I2C , you can connect all you want, or CAN bus . if you mean a proper Lidar, one device can usually scan all around, but then you need some processing power to analyze data and use them for control.
-
1
-
2
of 2 Next